Developing voice climates : does transformational leadership affect employees speaking up?

(1)

Developing Voice Climates

Does Transformational Leadership affect employees

speaking up?

Luuk van Balkom 10077316 Thesis Seminar Business Studies Supervisor: Renske van Geffen Academic year: 2013-2014 Semester 2, Block 2

Amsterdam, June 5th 2014

(2)

Abstract

Previous research agrees on the fact that the upward communication of ideas can have significant implications for an organisation’s performance and even its survival. Both voice climate and transformational leadership seem to have a positive influence on proactive behaviour and voice. For managers it is important to know whether to invest in creating a stimulating voice climate, attracting or training transformational leaders or both. This research adds to the stream of literature suggesting that employees feeling they can effectively voice tend to voice more often. However, this study found that transformational leadership does not necessarily have an effect on an organization’s voice climate itself. These results indicate that voice climate and transformational leadership may be interchangeable and seen as separate variables influencing employee voice. These results may be useful for managers who need to make a trade-off between promoting a positive voice climate and obtaining a transformational leader when looking to increase employees speaking up within their organization.

(3)

1. introduction

One way companies nowadays try to secure a sustainable competitive advantage is through stimulating their employees to show personal initiative and engagement. An example of this is Google’s “20 percent time” rule that enables employees to spend one day per week on projects that they are not necessarily required to do. One employee at Google, Alex, was reading articles using Google Reader when he found he wanted to skip several sections of stories but could not find any shortcuts on his keyboard. Alex decided to ask the Reader team whether they could introduce the feature and was told he could do so himself. When he found out he could, he did so right away and now admits he now feels thrilled every time he uses Google Reader (Google's "20 percent time" in action, 2006).

This Google example shows that a stimulating voice climate can lead to increased employee voice, which in turn, can lead to enhanced products. As organisations are continually shifting responsibility downward and require employees to go beyond their task descriptions, among other things they require their employees to voice and show initiative (Crant, 2000). In several recent studies (Van Dyne & LePine, 1998; Detert & Burris, 2007) voice is described as the improvement of organizational functioning through the communication of opinions, suggestions, ideas, or concerns. According to Detert & Burris (2007), as voice challenges the status quo, there is some risk involved for the actor. Because of this, individuals in organisations weigh the potential risks and rewards of certain voice behaviour before engaging in it.

They continue by arguing that voice climate is an important influence on these calculations. This voice climate refers to employees’ collectively held beliefs about the level of safety and efficacy of speaking up (Morrison & Milliken, 2000). The

(5)

former constitutes employees having to feel safe before they are willing to speak up. Morrison & Milliken (2000) comment that employees often believe they may be punished for raising sensitive topics. The latter, voice efficacy, will be high when employees expect they can communicate effectively and think their ideas will be taken seriously and acted upon. This study expects that both voice climate efficacy and voice climate safety have a positive effect on employee voice.

Furthermore, previous research suggests that certain leadership styles may influence employee voice and the favourability of the voice climate within an organisation. Morrison, Wheeler-Smith & Kamdar (2011), in their research on voice climate in groups suggest that more research is needed to investigate how voice climate develops. They note that leadership behaviours are likely to play a particularly important role. Morrison (2011) further supports this by arguing that leaders can send certain signals about the likely consequences of voicing in an organisation. In order to retain financial performance an organisation needs to find out what leadership style has a positive effect on an organisation’s voice climate and consequently on its employees’ voice behaviour.

One type of leadership previous research has focused on is transformational leadership. The key of transformational leadership is that it infuses work with meaning, expresses an attractive future vision, but most of all inspires followers (Den Hartog & Belschak, 2012). Because transformational leadership inspires followers this study believes it may have a vital influence on voice climate. However, as mentioned before, previous research has not yet been implemented on voice climate. The present study, therefore, will focus on the effect transformational leadership has on the favourability of an organisation’s voice climate and subsequently on the amount of voice presented by employees.

(6)

The goal of this paper is to investigate this gap by asking the following research question: Does transformational leadership have a positive strengthening effect on the positive relationship between a stimulating voice climate and employee voice?

To come to an answer to this question a survey will be held among Dutch speaking full-timers. With the results of the survey, an overview will be given to show how different employees value their organization’s voice climate and the transformational characteristics in their leaders, which might make it easier for supervisors to determine whether they should improve their organization’s voice climate, train or attract transformational leaders or do both.

This research will continue with a literature review, which will explore the different concepts through previous research. A conceptual framework will be presented which will contain the hypotheses. This is followed by the research design and methodology and eventually the results of the survey will be presented. The results will be addressed in the discussion and the paper will end with a brief conclusion.

+

Voice Climate Employee Voice Transformational

Leadership

(7)

2. Literature review

2.1 Voice

Several extant studies suggest that the upward communication of ideas can have significant implications for an organisation’s performance and even its survival (Morrison, 2011, Detert & Burris, 2007, LePine & Van Dyne, 1998). In

hypercompetitive markets it is increasingly more difficult for companies to gain and retain a competitive advantage. Employees are often in the possession of knowledge important to top managers for a company to respond adequately to the dynamic business conditions at hand. It is not, however, self-evident that this information reaches these top managers and therefore employee voice is a vital instrument in ensuring that it does.

Numerous authors have defined voice over the years. For an employee to be using voice in an organisation he has to comply with several characteristics. The first considers the content of voice itself. It has to be challenging yet constructive and is about expressing (rather than withholding) work-related relevant ideas, information, and opinions, suggested or needed changes (Tangirala and Ramanujam, 2008; Detert & Burris, 2007; Van Dyne & LePine, 1998; Premeaux & Bedeian, 2003).

The second characteristic is the target of voice. It has to be aimed at someone who holds power inside the organization in question. For voice to be effective it has to be directed at someone in the organization with who is perceived to have the authority to act upon this information. Voice needs not necessarily be aimed at a supervisor or senior manager but can also be aimed at members of one’s team (Morrison, 2011). Detert & Burris (2007) then, use these characteristics and define voice as information provided to someone inside an organisation with the intention to improve

organizational functioning.

(8)

Since voice is about individuals acting to improve a certain situation it is an example of proactive behaviour (Crant, 2000). Grant & Ashford (2008) support this view by arguing that voice is an individual anticipatory action taken by employees with the intent to impact themselves and their working environments. Furthermore, supervisors need not necessarily demand this type of voice behaviour. It can be classified as a type of non-required behaviour which adds to the fact that it is an example of voluntary proactive behaviour (LePine & Van Dyne, 1998). To conclude, voice is a voluntary individual anticipatory action taken by employees in an

organization with the intent to affect either their working environment or themselves. When put simply voice can be divided in three parts. These parts can be seen as constructive voice, warning others, or suggestion (Morrison, 2011). Voice content as described here is very broad. Premeaux & Bedeian (2003) argue that voice may be a view that differs from those of others in the organization. The message that is communicated can simply be a view that differs from the views of others in the organization, in which case the conveyer of the message is just trying to point out a difference of opinion to others (Premeaux & Bedeian, 2003). An employee can, however, also speak up to bring about change to certain work-related issues . (Milliken et al., 2003). In that case the employee has the intention to try to improve this situation either by himself or through his superiors. Last, voice can be a warning. This can mean that an employee sees issues that can either be harmful to the

organisation or to himself or his co-workers (Morrison, 2011). In the first case the employee witnesses an issue that is potentially damaging to the company and tries to advert damages by warning a supervisor. In the second case, considering himself or his co-workers, an employee may report a situation of misconduct or unfairness to his supervisor (Pinder & Harlos, 2001).

(9)

However, it turns out that employees often do not speak up. This is known as employee silence (Morrison & Milliken, 2000). Morrison & Milliken (2000) say this means that employees experience shared beliefs with their co-workers that speaking up to their supervisors with information about problems or suggestions is dangerous and futile. These shared beliefs are often caused because there is a difference in power between employees and their supervisors (Saunders et al., 1992). This implies that subordinates need to consider how their supervisors will react to them speaking up before they voice. Therefore, voicing can be seen as potentially risky. According to Liu, Zhu & Yang (2010) voice behaviour has three potential risks associated with it. These being that voice may be associated with discomfort, employees may gain a negative public label or image, and voice may damage the relationship of the employee with others in the organization and thus destroy social capital (Adler & Kwon, 2002). Because of this, employees tend to weigh the potential risks and

benefits of engaging in voice behaviour (Detert & Burris, 2007). If they sense the risk to be too high, employees are more likely to remain silent rather than speaking their concerns (Morrison, 2000). Therefore, it is evident that, since speaking up is

important and needed in organizations, a positive voice climate is desirable to encourage voice.

2.2 Voice Climate and voice

One of the contextual factors that may be of influence on voice is the phenomenon of voice climate. According to previous research voice climate is a relatively new concept that deserves further investigating (Frazier & Bowler, 2012). The concept is related to the climate of silence, a phenomenon put forward earlier by Morrison & Milliken (2000). This climate of silence refers to the fact that employees in a

(10)

significant amount of organizations apparently are aware of the truth about certain problems and issues in the organization but do not communicate this to their superiors. Following this, Milliken et al. (2003) conducted a study and found that perceptions that the organizational structure of culture does not support (upward) communication are the main incentives for employees whether to voice or not.

The concept of voice climate, too, is about employees’ perceptions of their surroundings. Voice climate is defined by Frazier & Bowler (2012) as employees’ shared perceptions of the extent to which they are stimulated to engage in voice behaviours. This is substantiated by Morrison & Milliken (2000) who argue that a voice climate refers to employees’ collectively held beliefs about the level of safety and efficacy of speaking up. A favourable voice climate is the opposite of a climate of silence in that it represents a climate in which employees have share beliefs that it is safe and worthwhile to communicate opinions, suggestions and concerns (Morrison, 2011).

To make sense of the environment in an organization employees interpret the characteristics of their work surroundings through social interactions (Berger & Luckmann, 1967). Because of this they develop certain shared perceptions of what constitutes acceptable behaviour. This includes whether they are encouraged to voice at work or not (Morrison & Milliken, 2000). These shared perceptions have been shown to have a substantial impact on following employee perceptions and behaviours (Naumann & Bennett, 2000).

These perceptions lead to two separate constructs put forward by Morrison & Milliken (2000). The first is voice safety and refers to the degree to which employees feel they have the safety to be speaking up without there being consequences for them doing so. Employees frequently believe they may be penalized for raising sensitive

(11)

topics. The second is voice efficacy, which refers to whether employees perceive their ideas and suggestions will be taken seriously and acted upon and whether they can communicate effectively in general.

Even though the work of Morrison & Milliken (2000) provides theoretical guidelines for the study of voice climate, little research has examined the issue further until recently (Frazier & Bowler, 2012). However, since speaking up poses a potential threat to an employee it is important to examine a climate that encourages voice further (Detert & Burris, 2007). Previous research has argued several contexts under which employees tend to voice more often in groups. Erez, Lepine & Elms (2002) found that when groups took on practices such as peer evaluations or certain styles of leadership, a positive effect on employee voice could be seen. This can make it more comfortable and less intimidating to voice and may also have a positive effect on an employee’s feelings of voice efficacy (Morrison, 2011). Therefore, some contexts can have a positive effect on how comfortable employees are with voicing and can

promote an organizational voice climate. Others, however, are less supportive and provide signals that voicing is likely useless and may even pose a threat to an

employee’s image or career, thereby supporting a climate of silence (Morrison, 2011). Morrison, Wheeler-Smith & Kamdar (2011) argued that when an individual comes to the point of deciding whether to voice or not, he or she will tend to be influenced by collective beliefs about the likely outcomes, and, thus, be influenced by the organization’s voice climate. Employees will be more likely to speak up when they receive social cues that suggest that others in the organization view speaking up as something that can be done effectively and safely.

On the one hand, these social cues may strengthen the employees’ collectively held beliefs that speaking up has no negative consequences and that their voice

(12)

reaches someone in the organization with the power to act upon it (Morrison, 2011). On the other hand, when employees receive social cues that supervisors are reluctant to change and employees tend to believe that voice is often punished or ignored they are less likely to contribute input when they have potentially valuable ideas to share (Morrison, 2011). Some organizational environments even consistently foster a climate of fear and intimidation that disapproves of speaking up to improve organizational situations (Pinder & Harlos, 2001).

Leaders in an organization can affect their employees directly through their control over organizational structures and policies and therefore have an impact on overall culture or climate (Detert & Burris, 2007). If these leaders promote a climate where their subordinates perceive it to be safe and effective to voice an environment is created in the workplace in which speaking up is acceptable and encouraged among employees. Because subordinates are looking for social cues in their environment, a climate promoted by supervisors that encourages the performance of voice behaviours by employees will probably lead to enhanced voice behaviours among employees (Frazier, 2009).

This study expects than that when an organization has a prevailing positive voice climate this will increase the amount of employees speaking up. Together, these arguments about voice climate considering the amount of employee voice suggest the following hypothesis:

Hypothesis 1a: A voice climate encouraging feelings of safety has a positive effect on the quantity of voice.

Hypothesis 1b: A voice climate encouraging the efficacy of speaking up has a positive effect on the quantity of voice.

(13)

2.3 Transformational leadership and voice

One of the many leadership styles covered in previous research is

transformational leadership. Previous research argues that the key to transformational leadership is the ability of the leader to inspire his employees to bring about

improvement and change within an organization (Bass, 1990; Den Hartog & Belschak, 2012). According to Bass (1990) leaders can be classified as

transformational when they are able to raise the interests of their subordinates, when they promote the mission and purposes of the group, and when they manage to make subordinates look beyond their own personal interest for the greater good of the group. This is substantiated by Den Hartog & Belschak (2012) who argue that transformational leaders can stimulate and encourage proactive behaviour because they focus on improvement and change. All in all employees are touched by their transformational leader because they associate with them. Employees are moved to act beyond their self-interests because transformational leaders themselves are often more change-oriented and proactive and, thus, they may act as role models to their subordinates (Crant & Bateman, 2000).

This stimulation of proactivity can be achieved in several ways. A leader has to possess several characteristics to be classified as transformational (Den Hartog & Belschak, 2012). Transformational leaders may act as role models to their

subordinates because they are commonly more proactive and change-oriented themselves (Crant & Bateman, 2000). Also, through increasing their subordinates’ powers and stimulating them intellectually (Den Hartog & Belschak, 2012) these leaders can arouse followers to a higher level of thinking (Bass, 1985). Furthermore, through infusing the work with meaning and the articulation of a captivating future

(14)

vision, these transformational leaders inspire their employees.

Bass (1985) argues that it is evident that charismatic leaders have great influence over their employees. Charisma is one of the characteristics a

transformational leaders should have because through being charismatic to their followers they may meet the emotional needs of their subordinates, intellectually stimulate their subordinates, and, once again, inspire them (Bass, 1985). This is substantiated by Bass (1990) who argues that intellectually stimulating a subordinate can lead to desirable outcomes. As employees want to identify with transformational leaders these employees may be stimulated with the idea that with some extra effort they are able to accomplish great things. Transformational leaders may show employees new ways of looking at old problems or by showing them to see

difficulties as issues to be handled, and by laying an emphasis on rational solutions (Bass, 1990). Also, transformational leaders pay close attention to differences among their subordinates and because of this can be seen as individually considerate (Bass, 1990). These leaders act as mentors to their subordinates and help them to develop and grow. Because of this employees have a high degree of confidence and trust in their transformational leaders and, therefore, these leaders appeal to their

subordinates’ values and ideals. In doing so, these leaders heighten the level of commitment of their employees and they stimulate their subordinates to formulate new ways of thinking about issues (Piccolo & Colquitt, 2006).

As argued before, these transformational leadership practices should lead to employees acting beyond their self-interests and them being change-oriented and proactive. Proactivity, then, is one of the desired outcomes of transformational leadership. Previous research suggests that certain leadership styles can either restrict or promote proactive behaviour (Crant, 2000). As mentioned before, voice is a form

(15)

of proactive behaviour so transformational leaders may have an impact on employee voice because they can directly stimulate proactive behaviour (Den Hartog &

Belschak, 2012).

Transformational leadership may then also encourage voice, since voice being an example of proactive behaviour. Detert & Burris (2007) argue in favour of this and say that transformational leadership encourages voice due to transformational leaders creating responsibility and commitment toward the greater good of the organization and encourage subordinates to become innovative issue solvers. In their research they found a positive relationship between voice and transformational leadership. Liu, Zhu & Yang (2010) and Bass (1985) corroborate this by arguing that the elements of transformational leadership, which they identify as inspirational motivation,

individualised consideration, idealised influence, and intellectual stimulation, should all stimulate subordinates’ voice behaviour. In their research they also proposed that transformational leadership gives rise to higher identification of employees with their supervisors and, therefore, more upward voice (Liu, Zhu & Yang, 2010). They also proposed that transformational leadership relates to employees’ higher identification with the firm, which leads to more voicing to co-workers.

The intimidation employees sometimes associate with speaking up are likely to be reduced by coaching (Detert & Burris, 2007). Coaching and vision sharing by leaders may also increase employees’ initial motivation to voice because they cause employees to accept more joint responsibility for an organization’s performance outcomes (Bennis & Nanus, 1985). These arguments lead to the fact that

transformational leadership practices tend to lead to increased employee

empowerment and felt responsibility to positively contribute to an organization’s future as well as to employee commitment and competence (Senge, 1990). These

(16)

employees should be more determined to speak up when they are convinced that their supervisors are oriented towards the future, rather than preserving the status quo (Detert & Burris, 2007). Transformational leadership seems to have a positive influence on proactive behaviour and this proactive behaviour should lead to an increased level of voice, this leads to the second hypothesis:

Hypothesis 2: Supervisors expressing traits of transformational leadership positively influence the amount of subordinates voicing.

2.4 Interaction effect of transformational leadership on the positive relationship

between voice climate and employee voice

As argued in paragraph 2.2, this study expects to find a positive relationship between an organization’s voice climate and the quantity of voice within an

organization. Hereafter it was argued that transformational leadership supposedly has a positive effect on employee voice. With these two hypotheses combined, this paragraph will focus on the interacting role of transformational leadership. More specifically, it will be argued that transformational leadership has a strengthening effect on the positive relationship between a stimulating voice climate and employee voice.

Morrison, Wheeler-Smith & Kamdar (2011) argued that leader behaviour and leadership style could play a significant role in the development of voice climate. This is due to the fact that leaders have the ability to send certain signals about the probable consequences of speaking up. They also say, however, that these arguments have not been empirically tested.

Zohar & Tenne-Gazit (2008) argue that an organization’s climate strength may

(17)

be affected by transformational leadership due to a number of reasons. These reasons involve the issues of leader-member relations and the expression of consistency across situations. These two issues both have to do with transformational leadership. First, considering leader-member relationships, they argue that transformational leadership has several attributes that boost a closer relationship with subordinates. Bass (1990) comments on this by saying that this relationship is characterized by a small power distance and that a transformational leader individually considers the needs and capabilities of his or her subordinates. This relationship is sustained by positive and encouraging verbal communication between leaders and subordinates and by mutual trust and openness (House & Shamir, 1993). Because of this, a stronger climate develops due to the fact that subordinates are better informed on what these leaders prioritize, support, and value (Zohar & Tenne-Gazit, 2008).

Second, the leadership practices exhibited by transformational leaders should exhibit greater consistency across situations (Zohar & Tenne-Gazit, 2008). Voice climate depends to a great extent on the perceptions of employees, referred to earlier in the theoretical framework. The variation in these perceptions may be reduced as leaders’ practices become more consistent. Because transformational leaders usually rely on visions and values when tackling issues, these leaders are expected to show more consistency. Also, this exhibition of increased consistency is especially true when subordinates’ safety is at stake (Zohar & Tenne-Gazit, 2008). This connects to an organization’s voice climate which is also, amongst other things, related to

subordinates’ shared perceptions of safety. Therefore, a transformational leader could enhance an organization’s voice climate when his or her subordinates’ welfare or safety is at stake.

(18)

Two variables that are especially important considering voice climate improving employee voice are work group identification and work satisfaction (LePine & Van Dyne, 1998). Employees satisfied with their working environment feel a stronger moral obligation to contribute ideas to a group and are, therefore, more motivated to voice. Also, employees that are highly identified perceive a strong connection with other members in their working environment and are more likely to contribute to the group through sharing ideas, recommendations, and opinions (Blader & Tyler, 2009, in Zohar & Tenne-Gazit, 2008). This connects to the arguments put forward earlier considering leader-member relationships. Transformational leadership has a positive influence on work group identification and work satisfaction and may, thus, positively influence an organization’s voice climate to improve employees’ voice behaviour. The previous arguments lead to the third hypothesis:

Hypothesis 3a: Transformational leadership has a positive effect on the positive relationship between voice climate efficacy and employee voice.

Hypothesis 3b: Transformational leadership has a positive effect on the positive relationship between voice climate safety and employee voice.

(19)

3. Research design & Methodology

After discussing previous literature and presenting a conceptual framework with hypotheses this section will focus on the methods how to empirically test these hypotheses. First the research design, which is a survey, will be discussed. After this the sample that is used is discussed followed by the measured that are used in the survey.

3.1 Research design

To test the hypotheses proposed in the conceptual framework a survey, or more specifically a questionnaire-based sruvey, is used. As this study uses existent literature to formulate a research question and theoretical propositions to formulate a framework to help the analysis, this study uses a deductive approach (Saunders, Lewis & Thornhill, 2009). A proper research method associated with a deductive study is the use of a survey. This survey will use a questionnaire because a lot of respondents are needed in a rather short period of time. Also, because the data obtained from a questionnaire are standardised they are easy to compare (Saunders, Lewis & Thornhill, 2009), which suits the research question. Easy comparison is also exactly what is needed to test the hypotheses on a large number of people. Another advantage is that the data collected can be analysed using descriptive statistics, which is a good way of analysing data when I want to compare a lot of different respondents from various firms. The hypotheses proposed in this research consider people with differences in leadership style and their subsequent effect on voice climate and different types of voice. To provide reliable results in testing the hypotheses, a large number of respondents is needed and this is exactly what a survey does. A survey ensures that reaching respondents is done quickly and affordably (Saunders, Lewis &

(20)

Thornhill, 2009). In addition to this, a survey is very useful for standardising the questions so every respondent consistently gets the same questions. This is done in order to compare the answers of different respondents and to make sure these comparisons are reliable (Saunders, Lewis & Thornhill, 2009).

To conclude, the questionnaire will be self-administered due to the fact that this is the fastest and easiest way to fill out the questionnaire since almost all questions have scaling and ranking types of answers. Also, because there is no interviewer present anonymity to the respondents is guaranteed and at the same time this reduces participant or subject bias (Saunders, Lewis & Thornhill, 2009). This will improve the general reliability of the data collected. One advantage of this study is that it uses quite a large number of participants and the fact that these participants come from a large variety of industries. Also, the use of a survey allows for the use of descriptive statistics because of the easy comparison of data through the standardised questions.

3.2 Sample

According to Saunders, Lewis & Thornhill (2009) the level of analysis depends on the goals of the study. Because the aim of this study is to find out whether transformational leadership has an effect on voice climate and employee voice in general the persons within an organisation will be taken as the level of analysis. Research will be focused on employees and their supervisors.

To help improve the level of generalization of the results data has to be collected from different sectors or firms (Saunders, Lewis & Thornhill, 2009). Therefore, this study has collected data from employees who are active in a number of different firms in the Netherlands. The focus of this study is on supervisors and

(21)

subordinates in different kinds of organizations. To be able to generalize findings a reliable sample had to be drawn (Saunders, Lewis & Thornhill, 2009). Because this study investigates different voice climates and whether or not leaders are transformational, a large number of participants, both employee and supervisor, was required.

According to Saunders, Lewis & Thornhill (2009) a minimum of 30 respondents is required in order for the data to be normally distributed which is required for most statistic testing to be carried out. A minimum of 200 respondents was desired for this research because a number of respondents were expected to drop out due to the extensive five-day length of the research. In practice the sample size was strived to be as large as possible due to the fact that results can be generalized better the larger the sample (Saunders, Lewis & Thornhill, 2009). Eventually, after deleting 105 respondents with missing values the final sample turned out to be one of 820 respondents.

3.3 Data collection

The method of sending questionnaires over the Internet was chosen rather than questionnaires distributed directly to respondents due to simplicity. A higher number of respondents can be reached this way and respondents get reminded every time they receive a questionnaire so the chance of them forgetting to fill one out decreases dramatically. The questionnaire was developed online by a thesis supervisor. Because no data have to be entered manually developing a questionnaire online provides the opportunity to make different types of questions, is easy to use of respondents, and saves both time and money (Saunders, Lewis & Thornhill, 2009). Respondents were approached by e-mail through my and some of my fellow students’ personal

(22)

networks. Therefore, convenience sampling and subsequently snowball sampling were primarily used for getting respondents. Although the convenience sampling may not fully represent the target population, this is slightly offset by the subsequent snowball sampling. A large number of respondents not known to the researcher from firms spanning different sectors replied with the desire to take part in the research, which contributes to a better representation of the target population. Because of this a large number of respondents from different geographical areas were able to be reached in a relatively short period of time. On the other hand, collecting data through convenience and snowball sampling may pose the threat of systematic sampling since people may choose whether they desire to participate in the research or simply ignore the email (Etter & Perneger, 1997). Systematic bias may be extended to the fact that only certain people respond. For one, respondents needed to have access to internet throughout the day, which not everyone has. Second, respondents needed to work from Monday through Friday. These people were surprisingly hard to find since most companies tend to never shut down nowadays which leaves a lot of employees working one or two days in the weekends.

The data will be collected by administering questionnaires to a sample of 210 Dutch-speaking full timers from different firms in several sectors. These participants work in various functions; employees and their supervisors may both be included. The research will be a diary study. Respondents first fill out a short questionnaire about their working environment which will take 10 minutes. After they are done with the preliminary questionnaire the diary aspect begins. During one working week

respondents have to fill out short questionnaires three times a day in the course of 5 days. The first when the respondents start their working day, the second when they are having lunch and the third when they are done with their working day. Each of these

(23)

questionnaires takes about two minutes. Filled out questionnaires go directly to the thesis supervisor so complete immunity is guaranteed and to thank respondents, a 50 euro gift card will be raffled among them. After the study is finished respondents will receive a practical overview of the results. A diary study was used because it poses several advantages over other types of studies. According to Scollon, Kim-Prieto & Diener (2003) a diary study, or experience sampling methodology (ESM) as they refer to it, has several major strengths. First, ESM increases the ecological validity of the study due to the fact that psychology is applied to real-life situations instead of in a laboratory. Second, disadvantages of traditional self-reports, such as memory bias, can be avoided because respondents are constantly reminded to fill in a questionnaire. This is supported by Czerwinski, Horvitz & Wilhite (2004) who argue that due to the fact that diary studies are carried out in the respondents’ real environment they have high ecological value. These strengths are a significant advantage to this study as it increases the validity and decreases bias associated with the research.

A flyer was sent beforehand to the respondents for them to know what the research was about. This flyer contained some basic information about the goal of the research without jeapoardizing the integrity of the research. Other information

considered what is being expected of the respondents, anononymity, and that they will receive an overview of the end results. The flyer is closed with information with which respondents can contact the thesis supervisor for questions, suggestions or other information.

The questionnaires themselves are all written in Dutch because it is this study is aimed at Dutch employees. The flyer and questionnaires both contain a University of Amsterdam logo with the aim to improve visual attractiveness and. This is done to

(24)

stress the importance of the research and with the hop eof improving the response rate (Saunders, Lewis & Thornhill, 2009).

3.4 Measures

The survey uses both behavioural variables and attribute variables to measure respondents’ work behaviour and characteristics, respectively. The measures used to analyse these variables will be described in this section along with the sequence in which they appear in the questionnaire. First the pre-test will be discussed, which contains questions concerning the attribute variables of transformational leadership. After this the daily surveys will be discussed which contain the behavioural variables concerning the voice behaviour of the respondents and their co-workers. A detailed copy of the survey is included in the appendix.

3.4.1 Pre-test

To measure the constructs put forward in the literature review and conceptual model some variables need to be operationalized (Saunders, Lewis & Thornhill, 2009). Voice climate will be the dependent variable and will be measured along two dimensions. The first dimension is voice climate safety beliefs which refers to the belief of employees whether speaking up is safe or dangerous. The second dimension is voice climate efficacy, which refers to whether employees perceive they can voice effectively. These dimensions are presented by Morrison & Milliken (2000) and in my opinion are the best representatives of voice climate in general.

Employee voice in general will be the independent variable. Voice will be measured through several things. These are the amount of improvement-oriented voice, the

(25)

sharing of ideas and the voicing of concerns. Transformational leadership will be the interaction variable.

The pre-test was administered in the weekend before the daily questionnaires were sent to the respondents. The pre-test consisted of questions about attribute variables which, amongst other variables, included transformational leadership and voice climate. Specifically, voice climate was divided into 2 sub-variables, namely, voice climate efficacy and voice climate safety. These variables were only measured once due to them being attribute variables. They included questions considering the employees’ general perceptions of their supervisor or working environment. These perceptions are established over time and do not differ from day to day.

The dependent variable Voice Climate will be measured using an adjusted seven point Likert scale with 12 items developed by Morrison, (2011). This and all subsequent scales are classified as adjusted due to the fact that they are English scales translated to Dutch. Voice climate efficacy and voice climate safety each have six items. An example of a voice climate efficacy item would be: “To what extent are people in your workplace able to effectively (effectively meaning the person at which voice is aimed actually acts upon it) communicate opinions about work issues to others in the team, even when the opinion is different and others may disagree?”. This scale, therefore, measures whether employees perceive their ideas and suggestions will be taken seriously and acted upon and whether they can communicate effectively in general. An example of a voice climate safety item would be: “To what extent are people in your workplace able to safely (safely meaning, without having to fear any negative consequences) speak up and encourage others in the team to get involved in issues that involve the team? This scale, then, refers to the degree to which employees

(26)

feel they have the safety to be speaking up without there being consequences for them doing so.

The other variable measured in the pre-test is the moderating variable Transformational Leadership. This will be measured using an adjusted seven point Likert scale with 11 items developed by De Hoogh, Den Hartog & Koopman (2004). An example of a Transformational Leadership item would be: “To what extent to you agree with the following propositions considering your supervisor: My supervisor gives employees the feeling that they are working together on an important mission/assignment”. This scale measures whether employees perceiver their supervisors to be transformational. When answering these questions about voice climate and transformational leadership, respondents need to answer on a scale from applicable to not applicable.

3.4.2 Daily questionnaires

The daily questionnaires measure will measure, amongst other things, whether employees voiced to one another or to their supervisor. The independent variable Employee Voice will be measured using an adjusted seven point Likert Scale with 3 items developed specifically for this research. These three items refer to the three types of voice identified in the literature (Morrison, 2011). “Did you give a work related suggestion?” corresponds with suggestive voice, “Did you voice about ways to improve existing processes?” represents constructive voice and “Did you voice about ways to prevent harm to the organization?” corresponds with the warning of others.

(27)

4. Results

This section will cover the results of the analyses put forward in the previous methodology section. First, descriptive statistics are to be discussed, after which the reliability of the scales is shown through Cronbach’s Alpha. Finally, correlations are shows and a linear regression is performed to test the hypotheses.

4.1 Descriptive statistics and reliability

This section will provide some general characteristics of the sample used and a description of the data that was collected. The total number of respondents that were approached to participate in this study was 935. However, 475 respondents failed to fully complete both the pre-test and the daily questionnaires which makes the final sample cover 460 respondents. The final sample (N=460) consisted of 267 (58%) male and 193 (42%) female respondents. The sample shows a large number of people (40.9%) betweeen the ages of 20 and 30 which is probably due to the convenience (and subsequent snowball) sampling used by the students conducting the survey. The sample contains a significant amount of highly educated people with 223 (48.5%) and 150 (32.6%) respondents having an HBO and University degree, respectively. Work tenure in the sample ranges from zero years, so people who have just started at their job, to 46 years with a mean of 8.97 years (SD = 10.11 years). However, by far, most respondents have started to work for their organizations only recently with 30.9% working at their organization for a year or less. The descriptive statistics are shown in the tables below.

(28)

Table 1a: Age sample

<20 20-30 31-40 41-50 51-60 >60

Sample

(N=820) 2.0% 39.6% 12.8% 19.8% 21.1% 4.8%

Table 1b: Education sample

Secondary

Education MBO HBO WO

Sample

(N=820) 3.5% 15.4% 48.5% 32.6%

Table 1c: Tenure sample in years

0-1 2-5 5-10 10-20 20-30 >30

Sample

(N=820) 30.9% 23.2% 15.0% 12.4% 15.2% 3.3%

For the data to be useful and of use for analysis, they need to be realible. Cronbach’s alpha was used to test whether the variables were reliable as this is the most

commonly used measure of scale reliability (Field, 2009). For a Cronbach’s alpha to be considered good it needs to be over 0.8 but any value over 0.6 is sufficient (Field, 2009). Voice has a reliability of 0.602, Voice climate efficacy of 0.925, Voice climate safety of 0.962, and Transformational Leadership of 0.952. Table 2 shows the

Cronbach’s alpha for the four variables. Whereas the alphas of voice climate efficacy, voice climate safety and transformational leadership can all be considered good, the alpha of voice is just sufficient.

4.2 Correlations

First the correlations between the variables will be discussed before the hypotheses will be tested. Table 2 shows the correlations between the different variables. As expected it turns out that employee voice is correlated positively with voice climate

(29)

efficacy, though only weak, r (460) = 0.161, p < 0.01. This implies that employees of an organization who feel that they can effectively voice, tend to voice more often. However, unexpectedly, it turns out that voice climate safety is not correlated with employee voice, r (460) = 0.090, ns. This would imply that people who feel that speaking up is safe, do not necessarily tend to voice more often. The same unexpected result was found for transformational leadership which is not correlated with

employee voice either, r (460) = 0.079, ns.

Table 2. Descriptives, correlations between the variables (Cronbach's Alph on diagnonal)

M SD 1 2 3 4

1. (IV) Employee Voice 4,855 1,208 (0,602)

2. (DV) Voice Climate Efficacy 5,347 1,09 ,161** (0,925)

3. (DV) Voice Climate Safety 5,509 1,145 ,090 ,654** (0,962)

4. Interaction Variable TL 5,121 1,171 ,079 ,406** ,294** (0,952)

Note. N=460. * p<.05 ** p<.01. IV = Independent Varibable, DV = Dependent Variable, TL = Transformational Leadership, M = Mean, SD = Standard Deviation

5.3 Regressions

The next step in this research would be to carry out a regression analysis to test the main effects between dependent variable Voice and independent variables (Voice Climate Efficacy and Voice Climate Safety) and the interaction variable

Transformational Leadership, respectively.

(30)

Model 1 in table 3 shows the three predicted main effects for Voice Climate Efficacy (β = .172, p < 0.01, R² = .027), Voice Climate Safety (β = -.027, ns), and Transformational Leadership (β = .017, ns) on Voice. Only voice climate efficacy seems to have a significant effect on employee voice. With an explained variance of 2.7%, it shows that voice climate efficacy affects employee voice in a positive way. Therefore, hypothesis 1b is supported although this effect is limited. This indicates that people who feel their voicing makes a difference tend to voice more often. Contrary to what was expected, neither voice climate safety nor transformational leadership appear to have any effect on voice climate, falsifying hypothesis 1a and hypothesis 2.

Hereafter, two interaction variables were created, namely Voice Climate Efficacy*Transformational Leadership and Voice Climate Safety*Transformational leadership. Another linear regression analysis was carried out to test whether these interaction variables had a moderating effect on employee voice. Model 2 in table 3 shows the results and unexpectedly both Voice Climate Efficacy*Transformational Leadership (β = .152, ns) and Voice Climate Safety*Transformational Leadership (β = .123, ns) do not have an effect on employee voice. These effects falsify hypotheses 3a and 3b and suggest that transformational leadership does not strengthen the effect voice climate efficacy and voice climate safety have on employee voice.

(31)

R² = 0.027 (β = 0.017, ns) + (β = 0.172**) + (β = -0.027, ns) R² = 0.027 + (β = 0.011, ns) + (β = 0.098, ns) VCE VCS Subordinate Voice Quality Transformational Leadership VCE VCS Subordinate Voice Quality Transformational Leadership 31

(32)

Table 3: Regression Results of main (Model 1) and interaction effects (Model 2) of Voice Climate Efficacy, Voice Climate Safety, and Transformational Leadership on Employee Voice, respectively.

Voice (DV) Model 1 Model 2

Coefficient SE Beta Coefficient SE Beta

(Constant) 3,901*** 0,334 4,825*** 0,059

Voice Climate Efficacy ,191** 0,071 0,172** ,186* 0,078 ,167*

Voice Climate Safety -0,029 0,065 -0,027 -0,024 0,068 -0,022

TL 0,018 0,052 0,017 0,054 0,055 0,052

VCE*TL 0,007 0,055 0,011

VCS*TL 0,066 0,051 0,098

R² 0,027 0,027

Note: Dependent variable is voice, N = 460 *p <.05, ** p <.01, *** p <.001

TL = Transformational Leadership, VCE = Voice Climate Efficacy, VCS = Voice Climate Safety

(33)

5. Discussion

This section will discuss the results found in the previous section. Both supported and falsified hypotheses will be discussed along with the implications they have when linked to the existing literature. After this, the theoretical and practical implications will be presented and the section will be concluded with the limitations and several suggestions for future research.

5.1 Voice climate and Transformational leadership in relation to voice.

The main goal of this study was to examine the interaction role between transformational leadership and both voice climate efficacy and safety and employee voice, respectively. Specifically, it was tested whether transformational leadership moderates the relationship between voice climate and employee voice. Employee voice was measured at subordinate level, both self-conveying voice and voice recognized from colleagues. This was done because voicing behaviour between managers is different from voicing behaviour between managers and subordinates. The following research question was proposed: Does transformational leadership have a positive effect on the positive relationship between a stimulating voice climate and employee voice?

The first hypotheses proposed argued that a voice climate encouraging feelings of safety and efficacy has a positive effect on the quantity of voice in an organization. However, this research only found a significant positive relationship between voice climate efficacy and voice climate, whereas voice climate safety turns out to have no significant effect on employee voice. On the one hand this in line with Morrison & Milliken (2000) because it supports their views that a positive voice climate may contribute to the extent to which employees expect they can

(34)

communicate effectively and think their ideas will be taken seriously and acted upon. This also supports Morrison (2011) who argues that positive social cues may strengthen the employees’ collectively held beliefs that them voicing will reach someone in the organization with the power to act upon it. On the other hand it contradicts the findings of both Morrison & Milliken (2000) and Morrison (2000) on voice climate safety. They found that employees often believe they may be punished for raising sensitive topics and whereas they found that a positive voice climate is needed for employees having to feel safe before they are willing to speak up no such effect was found in this study.

Since this study uses respondents from multiple sectors in the Netherlands it seems that employees only need their voice to be effective and they do not seem to pay attention to whether their working environment seems safe to voice. This might be something that is typical to the Netherlands. When looking at Hofstede’s (1984) study on culture dimensions in countries he argues that the Netherlands scores low on power distance. He argues that the Netherlands is characterized by being independent, equal rights, superiors are accessible, power is decentralized and managers count on the experience of their subordinates. These characteristics all suggest that employees in the Netherlands may naturally feel safer than in countries that score high on power distance.

Hypothesis 2 was falsified as well suggesting that transformational leadership does not positively affect employee voice in an organization. This was rather unexpected as most of the previous research considering transformational leadership argues that it has a significant positive effect on proactive behaviour of which employee voice is an example (Den Hartog & Belschak, 2012; Detert & Burris, 2007; Liu, Zhu & Yang, 2010). This effect, though, could be attributed to the combination

(35)

of using perceptual data and the fact that the sample consisted of quite young respondents. Perceptual data meaning that these respondents may not recognize their leaders to be transformational because in the end, whether a leader is seen as transformational rests on what their subordinates perceive them doing or be like (Bandura, 1989). Bandura (1989) further states that these perceptions may sometimes be ‘objectively’ wrong. Also most respondents had a relatively short tenure. This is also a factor that may play a role because these are the employees that probably tend to voice less due to them not being in the organization for long.

Although beforehand it was expected that transformational leadership would have an impact on voice climate it may come as no surprise, mostly due to hypothesis 1a and 2 having been falsified, that hypothesis 3b was not supported. On the other hand, since hypothesis 1b was supported it might be a little unexpected that hypothesis 3a was not supported. These results imply that transformational leadership does not have the strengthening effect on the relationship between voice climate safety and voice that was expected. This again might find its origin in the arguments put forward earlier about transformational leadership and voice climate. People who already frequently voice may not need transformational leaders to help them express their concerns or suggestions. Furthermore, the literature suggests several other factors and variables that may have an influence on voice in an organization. Den Hartog & Belschak (2012), for instance, note that there other factors that influence proactive behaviour and voice. Notably they describe two variables; job autonomy and self-efficacy. With regard to job autonomy they argue that when an employee has high autonomy and is able to do his job successfully, this reduces that employee’s need for guidance from a supervisor. They argue the same goes for self-efficacy as when an employee is high on self-efficacy and possesses the skills or abilities needed

(36)

to successfully perform his job this, too, reduces the effect a leader has on them. Therefore, when employees score high on these factors, they may not need an attractive voice climate or transformational supervisors for them to speak up.

Moreover, future studies may have to divide transformational leadership in the two components it exists of, being the charismatic and empowering component. Charisma may be more associated with the safety part of an organization’s voice climate whereas empowerment might be with voice climate efficacy. When these two constructs are measured separately, other effects on voice climate and voice may be found.

Furthermore, another way to look at voice climate, transformational leadership and voice may be to turn the relationships around. It may be fruitful to investigate whether employees’ speaking up has a positive effect on an organization’s voice climate instead of the other way round which is argued in the present study. The constructs are related but the causality may be opposite. The same could be argued for transformational leadership and voice climate. Future studies may look into whether an encouraging voice climate may positively influence transformational leadership as opposed to the other way round which is argued here.

5.2 Theoretical implications

This research adds to the stream of literature suggesting that employees feeling they can effectively voice tend to voice more often. It supports Morrison & Milliken’s (2000) research suggesting that when voice efficacy is high, employees expect they can communicate effectively and think their ideas will be taken seriously and acted upon, and therefore tend to voice more often. Next to supporting the stream of literature about voice climate efficacy this study also adds some ideas about voice

(37)

climate safety and transformational leadership. Whereas previous research all argued in favour of voice climate safety (Morrison & Milliken, 2000; Morrison, 2011) and transformational leadership (Detert & Burris, 2007; Den Hartog & Belschak, 2012; Liu, Zhu & Yang, 2010) this study found no such results. This would either imply that these results suggest new findings or that these findings are limited to this sample and the work settings investigated here. This study suspects the latter due to arguments put forward earlier concerning the sample included and the cultural dimensions of the Netherlands. The fact that no effect was found of both voice climate safety and transformational leadership may be because of the relatively low power distance in the Netherlands. Therefore employees in the Netherlands may naturally feel safer than would employees in countries that score high on power distance. This could explain why in the Netherlands employees, to a lesser extent, recognize the need for their organization’s to have a positive voice climate or a leader with transformational characteristics. These factors considering the possible lack generalizability and implications will be further discussed in the limitations of this research.

Considering the interaction effect of transformational leadership on voice climate and employee voice this study also adds some ideas to the transformational leadership and voice climate literature. In general scholars tend to agree that both transformational leadership and a positive voice climate have a positive influence on voice. However, this study found that transformational leadership does not necessarily have an effect on the relationship between an organization’s voice climate and employee voice. These results indicate that voice climate and transformational leadership may be interchangeable and seen as separate variables influencing employee voice. Just as Den Hartog & Belschak (2012) put forward other variables, besides transformational leadership, that influence voice the results from study

(38)

suggest that transformational leadership does not have a strengthening effect on voice climate but can merely seen as an interchangeable variable. This might be the case because employees who recognize the positive characteristics of transformational leaders need not necessarily experience an encouraging voice climate. The same goes the other way round; When employees feel that it is safe and effective to speak up to others, they may not need to identify with, or be inspired by, their direct supervisor.

5.3 Practical implications

The aim of this study was to find out whether transformational leadership could influence the relationship between an organizations prevailing voice climate and the amount of voice in the organization. For managers this information is of great value as it may influence the way in which managers should be trained to act or behave. Also it should provide information on how to get employees in an organization to speak up, and provide suggestions, opinions, concerns, and ideas. Since a positive effect was found of voice climate efficacy on employee voice it seems of great importance for managers to support and enhance their correspondence of the fact that employees can communicate effectively and think their ideas will be taken seriously and acted upon. On the other hand, it turns out that in our sample voice climate safety is of lesser importance and this may not have to be emphasized as them most

important point when wanting to increase employee voice.

Considering the interaction effect of transformational leadership was not found in this study it seems that Dutch managers should pick one or the other. Although the effect of transformational leadership was not found in our example this could be due to other factors explained earlier. Previous research found that, in general, both transformational leadership and a stimulating voice climate may enhance voice (Den Hartog & Belschak, 2012; Detert & Burris, 2007; Morrison &

(39)

Milliken, 2000). Either strengthening the organization’s voice climate or finding or training transformational leaders should have priority and not trying to implement both.

5.4 Limitations and suggestions for future research

This research does have several limitations and these should be noted. First of all, as mentioned before, this study uses perceptual data. Leader and climate influences ultimately rest on what subordinates perceive their leaders to have done or their climates to have been like (Bandura, 1989) even though such perceptions can be "objectively" wrong. Respondents may not experience a positive or negative voice climate or experience that their leader communicates transformational characteristics. Therefore, future studies should research the relationship between transformational leadership, voice climate and voice from multiple perspectives and sources.

Also this study was performed among Dutch respondents and the results may thus not be generalized to other countries. This is, again, due to the cultural differences between the Netherlands and other, especially non-European, countries. For future research, if one does wish to compare these results it may be fruitful to look at the strengths of Hofstede’s cultural dimensions (Hofstede & Bond, 1984) for other countries.

Third, there are some limitations to the use of a diary study. Two of these are self-selection bias and motivation as most people find it less motivating to be involved for such a long period of time (Scollon, Kim-Prieto & Diener, 2003). Some refuse to participate and a lot of people drop out some way into the process, either because they dislike being interrupted all the time or because they just forget to fill out the questionnaires. To overcome this it was strived to obtain as large a sample as

(40)

possible. Eventually only 460 of the 925 people filled out all the questionnaires necessary for this study, making it not a truly random sample, but overall this is still a fairly large and representative sample. Due to snowball sampling many respondents work for the same company, are of young age and relatively highly educated due to the connection with university students. However, this limitation is reduced due to the fact that a total of seven persons started with obtaining respondents and eventually a fairly large sample was obtained after all. Furthermore, because the data collection through questionnaires is done at one single point in time, it proves difficult to measure changes within a certain population unless more surveys are done at different points in time (Saunders, Lewis & Thornhill, 2009). The current questionnaire, however, diminishes this limitation through the fact that the questionnaire is administered three times a day for five consecutive days. This gives a better view of how people behave in an organization since they may not voice one particular day, but they might just on another. Another limitation of a survey is related to the number of questions that can be included in it. This is due to the fact that respondents are often unwilling to fill out large questionnaires (Saunders, Lewis & Thornhill, 2009) However, this research only uses one pre-test of ten minutes and the following daily questionnaires only take two minutes each which encourages people to take questionnaire. Another possible drawback with respect to using a questionnaire is that there is no chance to alter the questions, as might be done with qualitative research (Saunders, Lewis & Thornhill, 2009). Therefore, to overcome these limitations, the questionnaire was carefully designed. A limitation of sending questionnaires over the internet might be that respondents are less inclined to respond to these questionnaires but this may be offset by the fact that these people do enjoy anonymity, which they do not when they have to send back hand-written questionnaires.

(41)

For future research several other variables may be taken into account. Examples include respondents’ age and education. Most of all, tenure may play an important role as people voice more often the longer they work in an organization as they feel more secure and know their working environment.

With respect to the testing of the data a relatively low Cronbach’s alpha was found for employee voice. This was due to the fact that employee voice only used six items and low Cronbach’s alphas are quite usual for variables that use under 10 items in the questionnaires. Future research may want to incorporate more items in their employee voice variable to receive a higher Cronbach’s alpha.

(42)

6. Conclusion

All in all it is evident from the literature that both voice climate and transformational leadership have a positive influence on proactive behaviour and voice although, besides for voice climate efficacy, no evidence for that was found in this study. This may be caused by the cultural characteristics of the Netherlands or the arguments put forward in the literature concerning other variables that affect voice such as voice efficacy and job autonomy.

First, this research adds to the stream of literature suggesting that employees feeling they can effectively voice tend to voice more often. However, this study found that transformational leadership does not necessarily have an effect on an

organization’s voice climate itself. These results indicate that voice climate and transformational leadership may be interchangeable and seen as separate variables influencing employee voice. These results may be useful for managers who need to make a trade-off between promoting a positive voice climate and obtaining a

transformational leader when looking to increase employees speaking up within their organization.

(43)

7. Bibliography

Adler, P. S., & Kwon, S. (2002). Social capital: Prospects for a new concept. Academy of Management Review, 27, 17−40.

Bandura, A. (1989). Human agency in social cognitive theory. American Psychologist, 44: 1175-1184.

Bass, B.M. (1985). Leadership and performance beyond expectations. New York: Free Press.

Bass, B., M. (1990). From transactional to transformational leadership: Learning to share the vision. Organizational dynamics, 18(3), 19-36.

Belschak, F. D., Den Hartog, D. N. (2010). When does transformational leadership enhance employee proactive behavior? The role of autonomy and role breadth self-efﬁcacy. Journal of Applied Psychology, 97, 194–202.

Bennis, W., & Nanus, B. (1985). Leaders: The strategies for taking charge. New York: Harper & Row.

Crant, J. M. (2000). Proactive behavior in organizations. Journal of Management, 26, 435 – 462.

Crant, J. M., & Bateman, T. S. (2000). Charismatic leadership viewed from above: The impact of proactive personality. Journal of Organizational Behavior, 21, 63–75.

Czerwinski, M., Horvitz, E. and Wilhite, S. (2004). A diary study of task switching and interruptions. In Proc. CHI’04, ACM Press, 175-182.

Detert, J. R., & Burris, E. R. (2007). Leadership behavior and employee voice: Is the door really open? Academy of Management Journal, 50, 869-884.

(44)

Erez, A., Lepine, J. A., & Elms, H. (2002). Effects of rotated leadership and peer evaluation on the functioning and effectiveness of self-managed teams: A quasi-experiment. Personnel Psychology, 55(4), 929–948.

Etter, J-F., and Perneger, T. (1997) “Analysis of Nonresponse Bias in a Mailed Health Survey,” Journal of Clinical Epidemiology, 50:10, pp. 1123-1128

Field, A. P. (2000). Discovering statistics using SPSS for Windows: Advanced techniques for the beginner. London: Sage

Frazier, M. L. (2009). Voice climate in organizations: A group-level examination of antecedents and performance. http://search.proquest.com/docview/

305084042?accountid=14615. (305084042)

Frazier, M. L., & Bowler, W. M. (2012). Voice climate, supervisor undermining, and work outcomes: A group-level examination. Journal of Management.

Google's "20 percent time" in action (2006). Retrieved 13 April, 2014, from

http://googleblog.blogspot.nl/2006/05/googles-20-percent-time-in-action.html Grant, A.M., & Ashford, S.J. (2008). The dynamics of proactivity at work. Research

in Organizational Behavior, 28, 3–34.

Hofstede, G., & Bond, M. H. (1984). Hofstede's cultural dimensions: An independent validation using Rokeach's Value Survey. Journal of Cross-Cultural

Psychology, 15, 417-433.

House, R. J., & Shamir, B. (1993). Toward the integration of transformational, charismatic, and visionary theories. In M. M. Chemers & R. Ayman (Eds.), Leadership theory and research: Perspectives and direc- tions (pp. 81–103).

Liu, W., Zhu, R., & Yang, Y. (2010). I warn you because I like you: Voice behavior, employee identifications, and transformational leadership. Leadership

Developing voice climates : does transformational leadership affect employees speaking up?