Getting Credit for Proactive Behaviour: Do Men Get
Punished for Frequent Voice?
MSc Thesis – Business administration – Leadership and Management Track University of Amsterdam
Name: L.L.G. Zijp
Student number: 10868461 Supervisor: Inge Wolsink
Second Supervisor: Hannah Berkers Date: June 29th, 2015
Statement of Originality
This document is written by Student Leonie Zijp who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Table of contents
Abstract ... 6
Introduction ... 7
Theoretical Frame work ... 10
The Importance of Employee Performance Ratings for Organizations ... 10
Extra-role and Proactive Aspects of Voice Influence Performance Ratings ... 11
Why Voice Predicts Performance... 12
Conceptualizations of Voice ... 13
How Voicing O ften Leads to Better Voice ... 14
The Importance of Voice Q uality ... 14
The Role of Voice Quality in the Voice Q uantity to Performance Relationship ... 15
The Effect of the Gender Stereotype ... 16
Research Model ... 18 Research Method ... 19 Procedure ... 19 Sample ... 19 Measures ... 19 Voice Q uantity ... 19 Voice Q uality ... 20 Performance ... 20
Control Measures ... 20
Analysis and Predictions... 21
Mediation Model ... 21
Moderation Model ... 21
Results Section ... 22
Pre-analysis ... 22
Hypothesis Testing ... 25
Voice Q uantity, Q uality and Performance Ratings ... 25
Voice Q uality as a Mediator ... 25
Gender as a Moderator... 28
Post-Hoc Analysis ... 29
Simple Mediation in the Female Condition ... 29
Simple Mediation in the Male Condition ... 31
Differences Between Manager and Colleague Rated Voice ... 33
Discussion ... 34
Key Findings... 35
Limitations and Future Research ... 39
Methodological Limitations and Suggestions for Future Research... 39
Theoretical Limitations and Suggestions for Future Research ... 40
Contributions ... 41
Conclusion ... 42
References ... 44
Appendix ... 51
Abstract
This study investigated if the amount of ideas, problems and suggestions voiced influences the quality of these messages, and how this eventually leads to influence
performance ratings. Moreover, the study investigated how gender affects this relationship. Based on data from 85 Dutch teams, made up of two employees and their supervisor, the findings showed that an employee who voices often also voices original and useful messages. Moreover, when messages voiced are highly original and useful, the employees will receive a higher performance rating. We also found indications for two different mechanisms that influence performance ratings. First of all, performance ratings can be improved by often voicing high quality ideas, and this is primarily the case for women. Secondly, when
employees voice often, their performance ratings may decrease. After examining gender, we found that female employees reach a higher performance rating through voicing high quality ideas often. Men however, run a risk when voicing often. On the one hand, their performance ratings are positively influenced by how often they voice high quality ideas, but their
“A father and son are in a horrible car crash that kills the dad. The son is rushed to the hospital; just as he’s about to go under the knife, the surgeon says, “I can’t operate—that boy
is my son!” (Barlow, 2014)
This riddle has been going around for several years, but many people are still unable to answer it correctly. A recent study showed that only 15% of the people that were tested came up with the right answer: the surgeon is the boy’s mother (Barlow, 2014). The riddle shows how people unconsciously have predispositions about the work men and women ought to do. Biases like this one are unfortunately still very present in the work environment, and it is just a small step to go from gender bias in job occupation towards gender bias in presumed appropriate work behaviour. The type of work behaviour that is seen as appropriate and necessary has changed over the years (Detert & Burris, 2007), and since gender bias is mostly present in high performance oriented jobs, there may be a significant difference in what is seen as appropriate and high performance behaviour (Lorber & Farrell, 1991). An
increasingly important type of work behaviour that is also seen as an important part of employee performance, and is presumably expected more from men than from women, is voice behaviour. This is the constructive communication of ideas, suggestions or opinions about work-related issues aimed at trying to improve the organization (Morrison, 2011). Since the business environment is becoming increasingly dynamic and competitive, ideas regarding improvements can no longer solely come from top management and it has become more important for employees to speak up (Detert & Burris, 2007). However, voicing ideas and speaking up is a proactive, challenging and changing type of behaviour (Morrison, 2014; Whiting, Podsakoff & Pierce, 2008), which is why we expect that this behaviour might be seen as more appropriate for men than for women.
One of the reasons why gender may be of particular importance for voice behaviour is because voice is not part of formal job description. Voice behaviour is often seen as a type of extra-role and proactive behaviour, since it is a way for employees to go beyond the
requirements of their job (Morrison, 2014). Kidder and Parks (2001) state that in areas where roles are not formally described, such as extra-role voice behaviour, an interaction between gender and presumed appropriate behaviour may be even stronger since behaviours are more prone to subjective interpretation. A subjective interpretation of behaviour suggests that since the appropriateness of the behaviour is open to interpretation, gender stereotypes may
influence the expectations of the people observing the behaviour. For example, according to certain gender stereotypes, the typical man is assertive, active and rational, whereas the typical female is passive, submissive and emotional (Locksley, Borgida, Brekke & Hepburn, 1980). These sex stereotypes indicate that when it comes to voicing ideas, opinions or problems, men are more expected to do so, and the behaviour may thus seem more appropriate for them.
Moreover, voice is susceptible to gender bias due to the proactive nature of the behaviour. Research on proactive behaviour has found that women were more inclined to engage in helping behaviour, whereas men were more inclined to engage in challenging behaviour (Kidder, 2002). Since voice is primarily a challenging type of behaviour (Whiting, Podsakoff & Pierce, 2008), it is more in line with the presumed appropriate behaviour for men. Not behaving in line with presumed behaviour can have important consequences, such as receiving a lower performance rating (Heilman, 2001). Since accurate performance ratings are crucial for maintaining job commitment and improving performance on an ongoing basis (Church, 1995), it is thus important to understand how stereotypes and presumed behaviour can influence the way men and women are rated. However, an inclusion of gender in the
extra-role and proactivity literature is still sparse. To address this gap, we will include gender into our study.
Moreover, research on the relationship between proactive behaviours and performance ratings has often focussed on the act of behaving proactively, rather than the actual content of that proactive behaviour (e.g. Morrison, 2014). Therefore, it is still unclear what actually causes proactivity to lead to a higher performance, and there is thus a need to look into a distinction in the construct (e.g. Morrison, 2011). This is also the case for voice behaviour, and studies are inconclusive on whether voice has merely positive or negative consequences. For example, studies found that often, voice behaviour does lead to better performance (e.g. Van Dyne & LePine, 1998; Thomas, Whitman & Viswesvaran, 2010). On the other hand, Siebert, Kraimer and Crant (2001) found that engaging in voice behaviour could harm the possibility of future promotions and salary increases. These studies mostly focus on how often the employee voices rather than the actual content of that voice (Morrison, 2014), which could be an explanation for the inconclusive results. Studying the content of voice might thus be essential for understanding the influence of voice on performance ratings, and to address this issue we will take a closer look at the content by making a distinction between voice quality and voice quantity. Here, voice quantity is the frequency by which is voiced, whereas voice quality is the originality and usefulness of that voice (Wolsink, 2013). We use voice quality as a measure of content, because the usefulness and originality of the ideas of the employee are important for an organization to perform better (Oldham & Cummings, 1996), and this part of proactive behaviour is likely to be seen as the primary predictor for
performance ratings.
All in all, by not only examining how often somebody voices, but also the quality of that voice, we address the lack of an inclusion of voice content in the literature. Moreover, by examining the influence of gender on the relationship between both voice constructs and
performance, we aim to establish a complete view of the influence of proactive behaviour on performance. Also, by examining voice behaviour, the influence of voice on performance ratings and the effect of gender, this research contributes to the literature by (1) making a distinction between the constructs of voice behaviour, namely voice quality and voice
quantity, (2) advancing the understanding of the influence of voice behaviour on performance ratings by examining the relationship between the two constructs of voice, and (3) adding to the knowledge on gender and gender stereotypes, the influence of gender on perceived behaviour, and the consequences of gender stereotypes for performance ratings. Taking our contributions into consideration, our research question reads: to what extent does an
employees’ voice quantity affect their performance, and how is this relationship influenced by voice quality and gender?
Theoretical framework
The Importance of Employee Performance Ratings for Organizations
According to Rotundo and Sackett (2002), job performance is the ‘actions and behaviours that are under the control of the individual and contribute to the goals of the organization’. In order to translate these behaviours and actions into ratings, Viswesvaran, Schmidt and Ones (2005) compiled a list of job performance rating dimensions including dimensions such as the quality of the job performed and the interpersonal competence of the employee. The outcome of such a rating influences decisions on, among others, reward allocation and termination (Ferris, Yates, Gilmore, & Rowland, 1985). Moreover, Greenhaus, Parasuraman and Wormley (1990) found that performance ratings can influence the
promotability assessments of the employee. It is thus in the interest of the firm and the individual employee to understand what influences performance ratings in order to prevent suitable employees being overlooked for promotion opportunities or even being let go.
Since most performance rating systems require managers to make a subjective
evaluation of their employees (Landy & Farr, 1980), it is important to identify which factors influence such a decision. According to Orr, Sackett, and Mercer (1989), one of the possible influences on job ratings are organizational citizenship behaviours. Organizational citizenship behaviour (OCB) is a type of behaviour that is oriented towards promoting organizational effectiveness, but that is not officially part of the job description (Organ, 1988). The literature divides two types of OCBs, namely helping and challenging (e.g. Williams & Nadin, 2012). In this paper, we will focus on challenging behaviour since this type of behaviour is becoming increasingly important for firm success and individual performance ratings due to the
progressively changing and dynamic business environment (Van Dyne, Cummings, & Parks, 1995).
Extra-role and Proactive Aspects of Voice Influence Performance Ratings One type of challenging OCB is voice behaviour, which is ‘the informal and discretionary communication by an employee of ideas, suggestions, concerns, information about problems, or opinions about work-related issues to persons who might be able to take appropriate action, with the intent to bring about improvement or change’ (Morrison, 2014; p.174). There are two aspects to this definition. First of all, voice behaviour is a type of organizational citizenship behaviour since both OCB and voice are not part of the job description and aimed towards improving the organization (Podsakoff, MacKenzie, Paine & Bachrach, 2000). Often, voice is described as a type of challenging citizenship behaviour since it challenges the status quo, and is aimed to promote a positive organizational change (Whiting, Podsakoff & Pierce, 2008). Since voice behaviour is not part of the job description it is also seen as an extra-role behaviour, indicating that voice is a way for employees to go beyond their formal job descriptions and put in a little extra effort towards the good of the
firm (Morrison, 2014). Secondly, voice behaviour is proactive (e.g. LePine & Van Dyne, 1998), which means that it is self-initiated, future-oriented and aimed at improving a certain situation (Morrison, 2011). Voice is thus OCB, extra-role and proactive in nature, and since both proactivity and OCB have been shown to increase performance ratings (Grant, Parker & Collins, 2009; Podsakoff, MacKenzie, Paine & Bachrach, 2002), voice may be an important predictor as well.
Why Voice Predicts Performance
There are several reasons why voice behaviour can be important for performance ratings. First of all, voice behaviour facilitates the changes that are needed in a dynamic work environment and managers may thus perceive it as an important aspect of job performance (Van Dyne & LePine, 1998). Since the work environment has become increasingly dynamic, the need for voicing to solve problems and provide suggestions for change has also increased. When an employee exhibits such important voice behaviour, he or she may be rewarded by the manager through an allocation of a higher performance rating (Van Dyne & LePine, 1998). Secondly, Allen and Rush (1998) found that when a manager perceives an employee to be committed to the company, this employee will likely receive a higher performance rating. Since organizational commitment is shown by a willingness to exert effort on behalf of the organization (Mowday, Steers & Porter, 1979), employees who voice their suggestions and ideas for solving problems in the firm can be seen as being highly committed to the firm. Therefore, organizational commitment shown through the exertion of extra effort via voice behaviour can increase the performance rating of the employee. This is supported by Burris, Detert and Romney (2013), who argue that managers may view voice as a sign of concern for the organization and its well-being, which they will reward with higher performance ratings.
Therefore, we state that when employees voice frequently, managers will give them a higher performance rating.
H1: Voice Behaviour is positively related to Performance Ratings of Employees
Conceptualizations of Voice
In order to understand why voicing leads to higher performance ratings we need to focus on the content of voice as well. However, literature on voice has predominantly
focussed on the act of voicing itself (Wolsink, Den Hartog, Belschak & Sligte, 2015). This is defined here as the frequency by which is voiced, and referred to as voice quantity (Wolsink et al., 2015). Previous studies that focused on the relationship between voice behaviour and performance (e.g. Van Dyne & LePine, 1998; Thomas, Whitman & Viswesvaran, 2010), and that included a measure of voice as the frequency of voice, have been inconclusive on
whether voice has merely positive or negative consequences for performance ratings. These inconclusive results could be a consequence of the absence of different distinctions within the concept of voice. Some distinctions have been made; for example, Liang, Fahr and Fahr (2012) made a distinction between promotive and prohibitive voice. When voice is promotive, it is aimed at improving existing work practices to eventually benefit the organization, and comes with suggestions for improvement. Prohibitive voice on the other hand is an expression of concern about existing practices or incidents that may harm the organization, but an
employee does not necessarily provide solutions to these problems. However, none of the distinctions made in the literature focus on how the quality of the voice varied, and we thus do not know what makes the voiced information beneficial to the organization and helps to effectively implement change (Wolsink et al., 2015). Therefore, we will distinguish between voice quantity and quality, where quality is ‘the discretionary constructive communication of
suggestions, problems or opinions that are both original and useful to the organization’ (Wolsink et al., 2015; p. 8). We expect that voice quality is the main cause for the increase in performance ratings following frequent voice for several reasons.
How Voicing Often Leads to Better Voice. First of all, we expect that the distinction between voice quantity and voice quality does not just differentiate between separate
constructs, but that these two constructs are also interrelated. We argue that people who voice often can increase their quality either by practice or by frequency: the more information you communicate, the more likely it is that one message will be important (Thompson, 2003; Osborn, 1953). However, people who voice little do not have this advantage, and they may rely on other abilities (such as perhaps intellectual ability) to reach high quality voice. Another way in which voice quantity can lead to voice quality is because of experiential learning. According to Lewis and Williams (1994), experiential learning is learning by doing, or learning from experience. It involves doing something, reflecting how this went and then developing new skills from this experience. Therefore, people who voice frequently have more experience with voicing their ideas, which helps them develop better skills in voicing, and eventually benefits their voice quality. Moreover, besides this self-reflection, an
employee who voices often also enhances their learning due to an increase in feedback from their manager (London, 2003). The increase in learning will benefit the quality of voice, making the ideas voiced more original and useful. All in all, we argue that people who voice often reach a higher voice quality than people who voice rarely.
H2a: Voice Quantity is positively related to Voice Quality
The Importance of Voice Quality. Secondly, since the impact of voice on performance ratings also depends on what the employee voices (Morrison, 2014), voice
quality is also important in establishing the influence of voice on performance ratings. A reason for this is that according to the norm of reciprocity (Gouldner, 1960), people will reciprocate others that have helped them in some way. For example, supervisors may become more successful through the helpful or useful suggestions of an employee (Whiting,
Podsakoff & Pierce, 2008). Since voice that is of a high quality will thus provide the
supervisor with more helpful suggestions and useful ideas for improvement, according to the norm of reciprocity, a supervisor will reward the voice behaviour by giving the employee a higher performance rating. All in all, it is thus especially the quality of voice that matters for performance ratings.
H2b: An Employee’s Voice Quality towards his/her supervisor is positively related to the Performance Ratings of this Employee
The Role of Voice Quality in the Voice Quantity to Performance Relationship. We thus expect that voice quality is both influenced by voice quantity, as well as an important determinant for performance ratings. As we stated before, studies on the voice – performance relationship have provided inconsistent results. It is therefore important to examine the conditions under which proactive behaviour leads to higher performance ratings (Grant & Ashford, 2008), and we expect that voice quality is the main contributing factor in this relationship. The right type of voice behaviour can have important benefits for the organization, such as increased organizational and group performance (Morrison, 2014). However, merely voicing will not always increase the performance within the firm. Namely, frequent voice can even lead to an increase in turnover (Burris & Detert, 2013). However, when voice behaviour is constructive, it can improve the functioning of a firm which will lead to benefits for both supervisors and other employees (Whiting, Maynes, Podsakoff &
Podsakoff, 2012). So, the effect of voice on performance ratings depends mostly on the quality and constructiveness of the voice, more so than the quantity. Voice behaviour can influence how an employee is evaluated, but this relationship seems to depend on what and how the employee voices. Presumably, mediating effects of content are strongest if voice is seen as highly qualitative. Therefore, the relationship between voice quantity and performance may depend on the quality of what was voiced. We argue that people who voice often receive higher performance ratings because they voice better ideas, problems and opinions than those who voice rarely. Thus, the voice – performance relationship is positively mediated by voice quality.
H2: Voice quality is the main aspect that influences performance ratings, and it is because of voice quality that voice quantity and performance ratings are positively related
The Effect of the Gender Stereotype
However, there might be conditions where the path between voice quantity and performance is not mediated by quality. Morrison (2014) stated that several authors have posed the view that the employee-level effects of voice depend on how voice behaviour is perceived by for example the manager. Different perceptions of behaviour could be the result of gender roles. Gender roles are the beliefs about the desired attributes and behaviours of men and women (Kidder & Parks, 2001). Eagly (1987) states that according to gender roles, women should show a concern for others, whereas men should be proactive and show self-confidence. People are expected to behave in ways that are consistent with their gender role, and they may experience negative outcomes when they do not behave according to their role (Eagly, Karau & Makhijani, 1995).
The effect of gender roles on perceived voice behaviour can be twofold. On the one hand, voice behaviour is proactive and is therefore more related to the male gender role than the female gender role (Kidder, 2001). Moreover, Kidder (2001) found that the performance of OCB-civic virtue behaviour, which is also identified as voice behaviour (Graham, 1991), was strongly influenced by gender, and men were more likely to perform such behaviour. We can thus state that voice behaviour might be more consistent with the male gender role, and hence more appropriate for men than for women. The relationship between voice quantity and the perceived voice quality may therefore be stronger for men, because it is presumed by supervisors that men who voice often are performing their expected behaviours and are seen as having something useful to say, whereas women who voice a lot may be overstepping their boundaries and are creating problems.
On the other hand, since it is not expected from women to voice, they may receive credit for it (Kidder & Parks, 2001). Voicing can be seen as an extra-role behaviour for women, and when women voice, managers and colleagues perceive that the female is going beyond the requirements of her job. This increases the perceived quality of the voice, because if women go beyond the requirements of their job, they must have something useful to say. Men who voice however, are merely acting as required, and are therefore not necessarily voicing more high quality ideas. Therefore, we expect that the relationship between voice quantity and the perceived voice quality is stronger for women than for men. All in all, we thus have two competing hypotheses.
H3a: The relationship between Voice Quantity and Voice Quality is moderated by gender, and when the subject is male, the relationship is stronger and thus more positive. When the subject is female, however, the relationship is weaker and thus less positive.
H3b: The relationship between Voice Quantity and Voice Quality is moderated by gender, and when the subject is female, the relationship is stronger and thus more positive. When the subject is male, however, the relationship is weaker and thus less positive.
Research model
Voice
Quantity
Gender
Voice
Quality
Performance
Rating
1 demogra phi c i tem, ra ted by empl oyees , ma l e/fema l e 5 i tems on a s ca l e from 1 – 10, ra ted by the s upervi s or, fromAs hford a nd Bl a ck (1996) 14 7-poi nt Li kert
i tems , ra ted by s upervisors a nd col l eagues, from Wol s i nk, Den Ha rtog, Belschak a nd Sligte (2015) 26 7-poi nt Li kert i tems , ra ted by s upervi s ors a nd col l eagues , from Wol s i nk, Den Ha rtog, Belscha k a nd Sligte (2015)
Research method Procedure
In order to test our hypotheses, we collected multi-source data from 85 native Dutch speaking teams in a large variety of companies who worked together on a daily basis. Data were
collected over a period of three months, and collection took place during office hours with a researcher present. Teams consisted of two subordinates, who rated each other on voice quantity and voice quality, and one supervisor, who rated both the employees’ voice quantity, quality and performance.
Sample
After discarding one team that was invalid due to inaccurate information, we reached a final sample of 170 individual employees from 85 teams. 52.4% of our sample consisted of male employees, whereas 47.6% was female. The ages in our sample ranged from 18 to 64, with an average of 32.4 years (SD=12.72). Employee tenure varied greatly ranging from less than a year to 45 years, with an average of 6 years (SD=8.67). Since the population of the Netherlands is 50.49% females and the average age is 41 years old (statline.cbs.nl, 2014), we feel our data are generalizable for the Dutch population. Moreover, 38% of our sample consisted of female managers, whereas 62% was male.
Measures
Voice quantity. Voice quantity of the employee was rated by both the manager and the colleague. We used a 14-item scale developed by Wolsink, Den Hartog, Belschak and Sligte (2015). The first 3 questions were open questions that asked to answer the amount of times the colleague voiced ideas, opinions or problems over the last month. The last 11 questions were measured on a 7 point Likert scale, ranging from ‘never’ to ‘very often’. An
example question is “how often does your colleague voice solutions to problems within the organization”. A high score on the items indicated that the employee voiced frequently.
Voice quality. Voice quality of the employee was rated by both the manager and the colleague. We used a 20-item scale developed by Wolsink et al. (2015). The statements were measured on a 7 point Likert scale, ranging from ‘not applicable at all’ to ‘very applicable’. An example item is “advice my colleague/subordinate communicates is useful”. A high score on the items indicated that the employee voiced useful and original ideas and suggestions.
Before and after the voice quantity and quality questions, the participants answered control questions to ensure they understood the difference between voice quantity and voice quality. An example of a control question is “the following questions are related to: the
amount of ideas, suggestions and opinions my colleague gives or the quality and usefulness of the ideas, suggestions and opinions my colleague gives”.
Performance. Managers rated employee performance on 5-item scale, which was developed by Ashford and Black (1996). The statements asked the manager to rate the employee on a scale from 1 – 10 according to their most recent job rating in comparison to their colleagues. The statements are “Reaching work goals”, “The quality of his/her job”, “Ability to get along with others”, “ Ability to work in a team” and “Ability to get work done in time”.
Gender. Gender was measured as part of the demographics part of the survey, and assessed with a fixed-response item (1= male, 2=female).
Control Measures. We also included several control variables into our analysis. First of all, we included the gender of the manager. Since we expect employee gender to be of influence in the model, it is likely to assume that the gender of the manager may also influence our results. Secondly, we included the age of the employee. Several studies have established an influence of employee age on performance ratings (e.g. Waldman & Avolio,
1986), and we thus need to control for this variable. Thirdly, Sturman (2003) found that employees who are employed at a company longer are better performers. Since tenure could thus influence our results, we also included it as a control variable. Lastly, since we also measure voice quantity as the amount of problems, opinions and suggestions an employee voices per month, a higher amount of hours worked per week can suggest more interaction and opportunity to voice. Therefore, we also controlled for hours worked per week.
Analysis and Predictions
To analyse the data, we used IBM’s SPSS statistics. Using a path analysis-based moderation and mediation analysis in PROCESS we tested the simple mediation model and the moderated mediation model (Hayes, 2012).
Mediation model. We predicted a positive direct effect of voice quantity on performance (H1). We thus expect that people who voice often will receive a higher performance rating.
We predicted a positive and direct effect of voice quantity on voice quality (H2a). Therefore, we expect that people who voice often voice better.
We expected a positive direct effect of voice quality on performance ratings (H2b). We thus expect that people who voice more useful and original ideas to colleagues and managers will receive a higher performance rating.
For the mediation model as a whole, we expected a stronger indirect relationship between voice quantity and performance ratings when the relationship is mediated by voice quality (H2). So, we expected that when employees voice often, they will reach a higher quality which will then lead to a higher performance rating.
Moderation model. We expected that gender moderates the relationship between voice quantity and voice quality (H3), and we expect the relationship to be stronger and thus
more positive when the employee is male (H3a), or stronger and thus more positive when the employee is female (H3b). Therefore, we expected that when the employee is female (male), and voices often, the ideas and problems she voices will be seen as less useful and original than those ideas and problems voiced by men (women).
Results Section Pre-analysis
Reliability tests showed that all scale items had an alpha exceeding 0.7, indicating that the reliability for each of the scales was good (Table 1). Next, we examined the correlations between the variables and the control variables (Table 1). Unexpectedly, we found no association between voice quantity and performance (H1) (r=0.074, ns). Therefore, an employee who voices often will not necessarily have a higher performance rating than an employee who does not voice often. Moreover, in accordance with H2a, voice quality and voice quantity are positively related (r=0.449, p<0.01), indicating that an employee who voices often is also rated as having more useful ideas. As expected in H2b, voice quality and performance did correlate positively (r=0.506, p<0.01), indicating that an employee who has a high quality voice is often rated as a better performer as well. These results did not provide support for hypothesis 1, but did provide initial support for our hypothesis 2a and 2b.
Furthermore, employee gender was unrelated to voice quantity, voice quality and
performance. However, there is a difference between the correlations for men and women (Table 2). Voice quantity was uncorrelated to performance ratings for either gender (M: r=0.014, ns, F: r=0.145, ns). Therefore, women and men who voice often are not also rated as better performers. Voice quantity was correlated to voice quality, and this correlation was stronger for women (M: r=0.336, p<0.01, F: r=0.546, p<0.01), indicating that when an employee voices often, they also voice high quality ideas. Moreover, women who voice often
have an even higher voice quality rating than men. Moreover, voice quality was related to performance for both men and women, but the relationship was stronger for men (M: r=0.545, p<0.01, F: r=0.481, p<0.01). So, men and women who voice high quality ideas also have a high performance rating, but this rating is even higher for men.
Table 1: Descriptives, Correlations and Cronbach’s Alpha (N=170)
Variable Mean SD 1 2 3 4 5 6 7
1. Voice Quantity (IV) 4.80 1.48 (0.74)
2. Voice Quality (MeV) 4.48 0.89 0.449** (0.96)
3. Performance (DV) 7.69 0.96 0.074 0.506** (0.83)
4. Employee Gender (MoV) - - -0.029 0.009 -0.024
5. Manager Gender (CV) - - -0.087 -0.003 -0.071 0.198*
6. Employee Age (CV) 32.38 12.72 -0.030 -0.078 -0.089 0.133 -0.201**
7. Tenure (years) (CV) 6 8.67 -0.105 -0.176* -0.211** 0.086 -0.204** 0.694**
8. Hours per week (CV) 28.41 12.25 0.182* 0.058 -0.085 -0.094 -0.227** 0.324** 0.220**
Note: **. Correlation is significant at the 0.01 level (2-tailed); *. Correlation is significant at the 0.05 level (2-tailed). DV = Dependent Variable. IV = Independent Variable. MeV = Mediating Variable. MoV = Moderating Variable. CV = Control Variable. Employee Gender; 1 = Male, 2 = Female. Manager Gender; 1 = Male, 2 = Female.
Hypothesis testing
Voice Quantity, Quality and Performance Ratings . The results of our regression analysis can be found in table 3 and 4. First, we employed a simple mediation model with our variables voice quantity, voice quality and performance to examine our first and second hypotheses. As depicted in Table 2, H1 was not supported, as there was no direct effect of voice quantity on performance (β=0.03, ns). Thus, an employee who voices often will not necessarily receive a higher performance rating. Model 3 showed support for hypothesis 2a, since there was a direct positive effect of voice quality on performance (β =0.591, p<0.01). Therefore, when and employee voices high quality opinions, suggestions or problems, this is likely to lead to a higher performance rating. Our hypothesis 2b was tested in model 2, and there was a positive direct relationship between voice quantity and voice quality (β =0.262, p<0.01) supporting H2b. This indicated that people who voice often voice better ideas than people who voice rarely.
Voice Quality as a Mediator. The results of the full mediation model are shown in model 3. Our results show a positive indirect effect of voice quality on performance (β Table 2: Descriptives and Correlations Gender (N=167)
Variable Mean SD 1 2
M F M F M F M F
1.Voice Quantity (IV) 4.85 4.77 1.48 1.47
2.Voice Quality (MeV) 4.47 4.48 0.85 0.92 0.336** 0.546**
3.Performance (DV) 7.69 7.65 0.92 0.99 0.014 0.145 0.545** 0.481** Note: **. Correlation is significant at the 0.01 level (2-tailed); *. Correlation is significant at the 0.05 level (2-tailed). DV = Dependent Variable. IV = Independent Variable. MeV = Mediating Variable.
voice quality and our H2 was thus supported. Therefore, employees who voice often receive a higher performance rating because their suggestions, opinions and problems are more original and useful. However, the direct effect of voice quantity on performance was negative in the mediation model (β =-0.128, p<0.01). This indicates that there is a competitive mediation (Zhao, Lynch & Chen, 2010). The lack of a zero-order effect of voice quantity on
performance ratings, the presence of a negative direct effect of voice quantity on performance in the mediation model and the presence of a positive mediator, namely voice quality, are indicators of the presence of a negative mediation path. The presence of an omitted negative mediator for the relationship between voice quantity and performance ratings explains the lack of a zero-order effect, since the effect is cancelled out by the positive and negative mediation path. Therefore, when people voice often, they may on the one hand increase their performance rating when the ideas and suggestions they voice are of high quality, but they may also decrease their performance rating via a second negative mechanism. Suggestions as to why this occurred will be elaborated on in the discussion section.
Table 3: Process analysis
Model 1 Model 2
Dependent variable Performance Voice Quality
Coefficient (B) SE t Coefficient (B) SE t
Constant 8.261** 0.459 18.003 3.278** 0.405 8.096
Voice Quantity 0.027 0.050 0.536 0.262** 0.045 5.867
Age 0.005 0.008 0.638 0.001 0.008 0.095
Tenure -0.028* 0.012 -2.433 -0.014 0.010 -.407
Figure 1: Model for simple mediation
Note: n=170; **. significant at the 0.01 level tailed); *. significant at the 0.05 level (2-tailed).
Gender Manager -0.322* 0.163 -1.978 -0.010 0.144 -0.072
Gender Employee -0.016 0.150 -0.108 0.025 0.132 0.189
R2 0.282* 0.465**
Note. N=160; **. significant at the 0.01 level (2-tailed); *. significant at the 0.05 level (2-tailed).
Table 4: Process Analysis – Mediation
Model 3
Dependent Variable Performance
Effect SE LLCI ULCI t
Indirect Effect
(Voice Quality) 0.155** 0.042 0.083 0.251
Direct Effect (Voice
Quantity) -0.128** 0.048 -0.222 -0.033 -2.661
Total Effect 0.027 0.051 -0.073 0.127 0.536
Note. N=160; **. significant at the 0.01 level tailed); *. significant at the 0.05 level (2-tailed).
Voice Quantity
Voice Quality
Performance Rating
β=0.262** β=0.591** β=-0.128**Gender as a moderator. To test whether gender influences the relationship between voice quality and voice quantity (H3), we used a moderated mediation analysis. The results of our moderated mediation are shown in table 5. Unexpectedly, as depicted in model 4, the moderation effect of gender was not significant thus not providing support for either H3a or H3b (β =0.155, ns). Being perceived as somebody who voices high quality problems,
suggestions and opinions does therefore not seem to depend on whether you are male or female. However, examination of the interaction effect indicated that the indirect positive effect of voice quality did differ between genders (Male: β =0.113; Female: β =0.204) with Bootstrapped CI’s not including 0, as can be seen in table 6. These results indicate that the indirect effect of voice quality is larger for women than for men, and that women will thus reach a higher performance rating faster than men when voicing often. We thus decided to analyse the influence of gender further, to examine why the indirect effect was so much larger for women. In order to do this, we used Baron and Kenny’s (1986) suggestion to run two separate analyses on the data.
Table 5: Process Analysis – Moderated Mediation
Model 4
Dependent Variable Voice Quality
Coefficient (B) SE t LLCI ULCI Constant 4.328** 0.719 6.018 2.907 5.749 Voice Quantity (VQ) 0.036 0.136 0.266 -0.232 0.304 Gender -0.728 0.447 -1.627 -1.611 0.156 VQ*Gender 0.155 0.088 1.761 -0.019 0.328
R2 0.482**
Note. N=160; **. significant at the 0.01 level tailed); *. significant at the 0.05 level (2-tailed).
Post-hoc analysis
In order to further analyse the effects of gender we performed two different regression analyses following Baron and Kenny’s split sample procedure (1986). First, we ran our simple mediation model with only female employees. Afterwards, we ran a second simple mediation analysis, including only the male employees.
Simple Mediation in the Female Condition. For the regression with solely female employees (Table 7 and 8), we found that there was a positive direct effect from voice quantity to voice quality (β =0.349, p<0.01), but no direct effect from voice quantity to performance (β =-0.0830, ns). Therefore, women who voice often have a higher quality of ideas, suggestions and problems but they do not necessarily also receive a high performance
Table 6: Process Analysis – Conditional Indirect Effects at Values of the Moderator
Dependent Variable Voice Quality
Gender Effect SE LLCI ULCI
Voice Quantity Male 0.113** 0.050 0.029 0.232
Voice Quantity Female 0.204** 0.054 0.116 0.336
Note. N=160; **. significant at the 0.01 level tailed); *. significant at the 0.05 level (2-tailed).
p<0.01). Women who have a high quality voice thus often receive a higher performance rating. Looking at the indirect effect, we found that voice quality is a positive mediator in the relationship between voice quantity and performance (β =0.195, CI: 0.070; 0.368). However, the direct effect of voice quantity on performance shows a non-significant relationship (β =-0.083, ns), indicating full mediation. Therefore, the amount of suggestions, opinions and problems voiced by women leads to a higher performance rating because they voice higher quality.
Table 7: Process analysis – Gender is Female
Model 1 Model 2
Dependent variable Performance Voice Quality
Coefficient (B) SE t Coefficient (B) SE t
Constant 8.524** 0.687 12.411 3.531** 0.552 6.400
Voice Quantity 0.112 0.081 1.375 0.349** 0.066 5.331
R2 0.364 0.597**
Note. N= 77; **. significant at the 0.01 level (2-tailed); *. significant at the 0.05 level (2-tailed).
Table 8: Process Analysis – Mediation – Gender is Female
Model 3
Dependent Variable Performance
Effect SE LLCI ULCI t
Indirect Effect (Voice Quality)
Simple Mediation in the Male Condition. We found the following results in the regression analysis with only male employees (Table 9 and 10). Voice quantity was positively related to voice quality (β =0.184, p<0.01), and voice quality was positively related to
performance (β =0.567, p<0.01). These results are similar to the results found for the women, and men who voice often or men who have a high voice quality will have a higher
performance rating. However, voice quantity was unrelated to performance in the zero-order effect condition in model 1 (β =-0.030, ns), and like in the condition with only women, this indicates that voicing often for men will not directly influence their performance rating. Moreover, in our mediation model, the indirect effect of voice quality was positive (β =0.104, CI: 0.018; 0.219), but the direct effect of voice quantity on performance was negative (β =-0.134, p<0.05). We have again found evidence for a competitive mediation (Zhao et al, 2010), and men who voice often receive a lower performance rating because of a negative mediation path that influences the relationship. However, on the other hand, men who often voice high quality problems, suggestions and opinions do receive a higher performance rating. This is unlike the female condition, where women could only increase their performance rating by often voicing high quality ideas, but voicing often would not have an indirect negative effect on their performance rating.
Direct Effect (Voice Quantity)
-0.083 0.087 -0.256 0.090 -0.958
Total Effect 0.112 0.0814 -0.050 0.274 1.375
Note. N=77; **. significant at the 0.01 level tailed); *. significant at the 0.05 level (2-tailed).
Table 9: Process analysis – Gender is Male
Model 1 Model 2
Dependent variable Performance Voice Quality
Coefficient (B) SE t Coefficient (B) SE t
Constant 7.753** 0.517 14.999 2.943** 0.511 5.761
Voice Quantity -0.030 0.060 -0.505 0.184** 0.059 3.115
R2 0.461** 0.459**
Note. N= 83; **. significant at the 0.01 level (2-tailed); *. significant at the 0.05 level (2-tailed).
Table 10: Process Analysis – Mediation – Gender is Male
Model 3
Dependent Variable Performance
Effect SE LLCI ULCI t
Indirect Effect (Voice Quality)
0.104** 0.050 0.018 0.219
Direct Effect (Voice Quantity)
-0.134* 0.053 -0.239 -0.029 -2.545
Total Effect -0.030 0.060 -0.149 0.089 -0.505
Note. N=83; **. significant at the 0.01 level tailed); *. significant at the 0.05 level (2-tailed).
Differences Between Manager and Colleague Rated Voice. In order to further analyse the presence of the negative mediator in the male mediation model, we performed a correlation analysis including a distinction between manager and colleague rated voice behaviour (Table 11 and 12). According to the correlations, women consistently have a larger association between voice quality and voice quantity (Manager ratings; F: r= 0.449, p<0.01, M: r= 0.264, p<0.05; Colleague ratings; F: r= 0.495, p<0.01, M: r= 0.279, p<0.05). This indicated that women who talk often were often also perceived as voicing useful and original ideas. This association was much weaker for men and when men voice often, they are not also perceived as having such a high voice quality. Moreover, we found that in the colleague rated condition, women’s voice quality did not correlate with their performance rating (r= 0.080, ns), indicating that women who voice useful and original ideas to their colleagues will not also have a high performance rating. On the other hand, in the manager rated condition, women’s voice quality was strongly related to their performance rating (r= 0.575, p<0.01), and women who voice useful ideas to their manager thus also have a high performance rating. Men’s voice quality was related to their performance rating in both the manager and colleague rated condition (Manager: r= 0.566, p<0.01, Colleague: r= 0.282, p<0.05). Therefore, men who voice high quality to their manager and to their colleagues also have a high performance rating.
Discussion
The goal of this study was to examine if there are gender differences in whether and how the frequency of employee voice results in favourable performance ratings. Specifically, we investigated if colleague and manager ratings of the usefulness and originality of
Table 11: Descriptives and Correlations gender MR (N=167)
Variable Mean SD 1 2
M F M F M F M F
1.Voice Quantity (IV) 4.66 4.87 2.03 1.77
2.Voice Quality (MeV) 4.48 4.42 1.05 1.14 0.264* 0.449**
3.Performance (DV) 7.69 7.65 0.92 0.99 0.043 0.103 0.566** 0.575** Note. N=160; **. significant at the 0.01 level (2-tailed); *. significant at the 0.05 level (2-tailed). Table 12: Descriptives and Correlations gender CR (N=167)
Variable Mean SD 1 2
M F M F M F M F
1.Voice Quantity (IV) 5.00 4.67 1.72 1.53
2.Voice Quality (MeV) 4.43 4.56 0.92 1.00 0.279* 0.495**
3.Performance (DV) 7.69 7.65 0.92 0.99 0.020 0.140 0.282* 0.080 Note. N=160; **. significant at the 0.01 level (2-tailed); *. significant at the 0.05 level (2-tailed).
employee voice mediated the relationship and whether mediation depended on gender of the voicer.
Key findings
Our prediction that voice quantity would increase the performance ratings of
employees (H1) was not supported. This indicates that managers do not give employees high performance ratings for merely voicing many problems, suggestions and opinions. Our finding differed from studies stating that there is a direct effect from voice to performance evaluations (e.g. Whiting et al., 2008). One explanation could be that voice only leads to higher performance ratings when the voiced messages are of high quality. Therefore, the quality of voice would be a logical mediator to explain previously found relationships between voice and performance.
We thus examined voice quality as a variable that could possibly explain the relationship between voice quantity and performance ratings. Firstly, our prediction that higher amounts of problems, suggestions and opinions voiced would lead to higher quality of these messages (H2a) was supported. This is in line with literature on both experiential
learning and frequency (Lewis & Williams, 1994; Thompson, 2003), stating that experience would lead to higher quality due to developed skills, and where more ideas would logically also include more high quality ideas (Goldenberg & Wiley, 2011). Therefore, employees can improve upon the quality of their opinions, suggestions and problems simply by voicing often. Secondly, we predicted that the quality of the problems, suggestions and opinions voiced would increase the performance rating of employees (H2b), which was supported by our results. Our result was in accordance with Morrison’s argument (2014), which states that the influence of voice on performance evaluations depends on what the employee voices. We
found that what the employee voices had to be both original and useful in order to lead to higher performance ratings.
The support for our previous predictions pointed towards the existence of a mediation effect of voice quality in the voice quantity - performance rating relationship (H2), which was supported by our analysis. This indicated that there was indeed a mediation effect, and when an employee often voices high quality, their performance ratings will improve. The finding of voice quality as a mediator shows the importance of voice quality for the voice - performance relationship, and shows that research into the content of voice, especially voice quality, is very important. Our findings addressed the results found in the literature on the positive and negative effects of voice on performance (Morrison, 2014) by showing that manager’s responses to voice depend on the quality of that voice.
Another point of interest in our study was the effect of gender. We predicted that due to gender roles and a difference in expected behaviours from men and women, the gender of the employee would influence whether frequent voice leads to higher quality ratings of voice (H3). However, this prediction was not supported, indicating that both women and men get higher quality ratings if they voice often. Yet, our results did show an interesting difference in the indirect effect sizes, where the indirect effect of voice quality in the voice quantity to performance rating relationship was much stronger for women than for men. We found that for women, performance was indirectly influenced by voice quantity, with a full mediation of voice quality. Thus, women who voice often get higher performance ratings because they voice better ideas, problems and opinions than women who voice rarely.
On the other hand, our results for male employees indicated that performance ratings were influenced by two separate mechanisms. First of all, performance ratings depended on the mediating effect of voice quality, where voicing often and voicing high quality would indicate a higher performance rating. This finding was similar to the mechanism for
performance ratings for women. Secondly, we found a direct effect from voice quantity to performance, which was not present in the female sample. The interesting finding here was that men who voice often actually decreased their performance rating via this direct
mechanism. So if frequent voice does not result in high quality voice, it can have detrimental effects on performance ratings. This was likely the result of possible omitted mediators (Zhao et al., 2010) and/or because the relationship between voice quantity and quality was weaker in the male sample. Thus, whereas for women, the effect of voice quantity on performance ratings is fully mediated by voice quality, for men, the effect is both positively mediated by voice quality and negatively mediated by another variable. This finding contradicts research stating that men are more likely to be rewarded for proactive behaviour than women (Kidder & Parks, 2001). We actually found the opposite, namely a negative possible mediation, indicating that men may actually be rewarded less for their proactive voice behaviour than women. Also, the positive indirect effect was much lower, indicating that it is easier for women to reach high voice quality and eventually a higher performance rating than it is for men.
We also found the negative effect of voicing often in our overall model, but after examining gender, it seemed that the negative mediator was only present for men. Our finding of a negative effect is in accordance with studies that showed that voice can also have a negative impact on performance ratings (e.g. Burris et al, 2013). Previous studies have examined possible negatively mediating variables that could influence performance ratings. An example is the study of Burris (2012), who showed that a change-oriented type of voice, where employees speak up in order to alter or destabilize accepted practices, is likely to decrease performance ratings. Moreover, Milliken, Morrison, and Hewlin (2003) argued that speaking up can lead to damaging your public image. According to Wayne and Liden (1995), image indirectly affects your performance ratings, and it could therefore also be a possible
negative variable influencing the voice quantity to performance ratings relationship. However, none of these studies included the quality of voice or looked at a difference between gender, and it is therefore not clear if these negative effects are present for both men and women.
There could be several other possible explanations for the presence of a negative mediator only in the male model. First of all, a common perception is that women talk more often than men (Broverman, Vogel, Broverman, Clarkson & Rosenkrantz, 1972). This could indicate that since it is less expected for men to talk, managers may be inclined to rate a male employee who voices often lower than a female employee, due to gender role incongruence (Eagly, Karau & Makhijani, 1995). Secondly, men are believed to be louder and more assertive (Broverman et al., 1972; Eagly, 1987), which could indicate that they are more prone to challenging and more negatively viewed types of voice (Burris, 2012; Kidder, 2002). This is supported by Tannen (2001), who stated that for men, conversation is a contest to achieve the upper hand or prevent others from pushing them around. For women on the other hand, conversation is held to exchange confirmation and support. Therefore, men might be more likely to engage in negative forms of voice, which may eventually decrease their performance rating. An example of more negative voice that might be used more often by men is prohibitive voice, which we discussed earlier. Men might be more likely to engage in prohibitive voice, since this type of voice is mostly used by highly assertive people (Liu, Liao, Liao & Wei, 2014; Kidder, 2002). Promotive voice however, might be practiced more by women, who are more likely to provide suggestions for improvement rather than calling out negative aspects. Thirdly, in line with our moderation argument, women may be more likely to be rewarded for voice because it is an extra-role behaviour in the female gender role (Kidder & Parks, 2001). Supervisors might give women credit for going beyond the expected requirements and thus give them a higher performance rating. Whereas men, of whom it is an
expected requirement to voice, are merely acting as they should and will thus not receive a higher performance rating.
We also examined the difference between manager and colleague rated voice, and found that women who voice high quality ideas and suggestions to colleagues are not also rated as high performers, whereas men are. This could indicate that when women have a suggestion or idea, they do not voice this idea to their manager directly but first to their colleagues. Voicing the idea first to colleagues gives them the advantage of being able to receive feedback on the idea and regulate the voice behaviour (Ashford & Tsui, 1991). Only when the colleagues respond well to this voice would women voice the same ideas to their manager. Therefore, the ideas they eventually voice to their manager will be of higher quality and thus more likely to have a positive effect on performance ratings. Men might not utilize their colleagues as a feedback mechanism and run the risk that what they voice to their manager is of such low quality that it is not accepted, which can reflect badly on their performance rating.
Limitations and Future research
Methodological Limitations and Suggestions for Future Research. Several limitations of our study should be taken into consideration. First of all, problems may have occurred with the generalizability of our data, since our study only focussed on the Dutch industry. However, our research did include a sample that was similar to the entire Dutch population, as we stated in the research method section of the study, and our sample spread across several different industries, which has increased the heterogeneity of our sample. But, our research is only generalizable to the Dutch population, and future research should expand the research area to different countries. Right now, our study would be more suitable for a generalization to other western or European countries, but not to for example Asian societies,
where gender stereotypes and expected work behaviours are different (e.g. Gelfand, Erez & Aycan, 2007).
Secondly, the cross-sectional attributes of our data leave us with only theoretical
assumptions of causality. These cross-sectional attributes of our data are problematic when we consider that the effect of voice quantity on voice quality may become stronger over time, due to increased learning and more experience. Moreover, there may be a tipping point where voice quantity may no longer influence voice quality and the employee is at a steady flow of high quality ideas. How this relationship changes over time, and how it may influence
performance ratings over time, should therefore also be examined in future research. In order to do this, researchers should use a longitudinal design. They should use a panel study, examining several teams over a longer period of time, while measuring voice quality, voice quantity and performance ratings fluctuations. Moreover, future research should focus on examining the longer term influences of voice, since manager’s perceptions of voice and the following performance rating may vary over time. Performance ratings may also be dependent on for example the development of a relationship between the manager and the employee over time, indicating a need to control and measure the role of time in our study. We did, however, ask managers and colleagues for the voice behaviour of the employee during the last month, which enabled us to look at an employee’s voice behaviour over a period of time. However, we were not able to look at how frequency influences quality over this period of time.
Theoretical Limitations and Suggestions for Future Research. Our results suggested that there is an omitted variable that is influencing the relationship between voice quantity and performance ratings negatively, for both the model including men and women and the model with only men. We did not find an indication for a negative mediator in the model with solely females. Future research should thus focus on finding possible variables that may be negatively mediating the relationship between voice quantity and performance
especially among male employees. We have already provided some possible variables, such as gender role incongruence and negatively viewed types of voice, that could explain how voice quantity could negatively affect performance ratings via a fourth variable, but these statements are merely suggestions and need to be examined.
Contributions
Theoretical Contributions. Our study contributes to the field of voice in several ways. First of all, this study has answered the call of Morrison (2011) for a content distinction in the voice construct. This study provides evidence for the existence of two separate
constructs of voice, namely voice quality and voice quantity. Moreover, we show that these constructs are interrelated, and that the quality of voice mediates the relationship between voice quantity and performance ratings. Therefore, our study indicates that the positive relationship found in the literature between voice and performance ratings are most likely the result of voice quality. Also, the study shows that the presence of two different mechanisms, namely a positive mediating effect of voice quality and an omitted negative variable effect, could have caused the differences in outcomes of previous studies that examined the effect of voice on performance ratings. Depending on the influence of the mediator, the relationship between voice quantity and performance ratings can be either positive or negative. Moreover, we have contributed to research on gender differences and the relationship of gender with voice. Kidder (2002) called for an inclusion of gender in the proactivity literature, and we have shown that gender is of influence in the voice – performance relationship. We found that men may face more negative consequences as a result of voicing than women, and researchers should be mindful of this difference when examining proactive behaviours in the future. This also indicates that the positive and negative results found in the literature on the relationship
between voice quantity and performance are likely the result of a difference in gender, where the relationship is negative for men but positive for women.
Practical Contributions. Our results suggest that people who voice often also voice useful and original contributions. It is thus beneficial for companies to create an environment where employees feel safe to voice their opinions, suggestions and ideas which will allow them to practice and gain experience, eventually leading to a higher voice quality. Since voice has been proven to lead to organizational effectiveness (Detert, Burris, Harrison & Martin, 2013), a free flow of high quality voice will thus be important for organizations as a whole. For example, leaders of teams should create psychological based safety by downplaying the power difference, which will encourage employees to speak up (Edmondson, 2003). Also, our research shows that speaking up can have important positive consequences for women. Their voice quality increases faster than that of men as a result of either voicing often, or only speaking up when they have something useful and original to say. This also causes women’s performance rating to increase faster. It is therefore important that women speak up, since an increase in performance ratings can have important long term consequences as well, such as promotions (Greenhaus, Parasuraman & Wormley, 1990). Men however, should be mindful when speaking up and take into account that when they voice, their voice should be of a high quality. When it is not, it may reduce their performance rating.
Conclusion
The aim of this study was to provide insight in how the frequency of voice could influence the quality of what people voice and how this can eventually influence an
employee’s performance ratings. We found that how often somebody voices could lead to a higher performance rating if that voice was also of a high quality. Moreover, we found that there was no difference between men and women in how voicing often leads to a higher
