Adventurous risk-takers contribute original voice,
independent of cognitive ability.
Name: Gizem Özmen
Student number: 10010823
University: University of Amsterdam
Faculty: Faculty of Economics and Business Section: HRM & Organisational Behaviour
Master Thesis: MSc. Business Administration – Leadership and Management Track First Supervisor: I. Wolsink, MSc.
Second Supervisor: R. E. van Geffen, MSc. Date: June 29, 2015
2 Statement of Originality
This document is written by me, Gizem Özmen, and I 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.
Foreword
I would like to thank my family and friends for supporting me throughout my academic career. I would like to thank my fiancé for always being there for me and making me believe in myself whenever I needed some reassurance. I would also like to thank my amazing supervisor Inge Wolsink for being available and helpful whenever I could not seem to get further in writing my thesis (both bachelor and master theses). I would like to thank all the participants for taking the time and effort to participate in this study. Last, I would like to thank you, the reader, for showing interest in my thesis and taking the time to read it. I hope you enjoy it and that it might help you get a better insight into the concept of employee voice behavior.
3 Table of Contents
Abstract ___________________________________________________________________ 4 Introduction ________________________________________________________________ 5 Theoretical Framework _______________________________________________________ 8 The Challenging Nature of Voice _____________________________________________ 8 Risky Situations or Risky People? ___________________________________________ 10 Employee Voice as Risky Decision-making Behavior ____________________________ 11 The Importance of Voice Quality for Organizations _____________________________ 12 How Risk-taking Propensity Might Lead to High Quality Voice Through Practice _____ 13 How Cognitive Ability Might Foster Voice Quality _____________________________ 14 Risk-taking is Good, But Only When Working Memory Capacity is High ____________ 16 Method __________________________________________________________________ 17 Design and Sample _______________________________________________________ 17 Measurements ___________________________________________________________ 18 Dependent Variable Employee Voice Quality ________________________________ 18 Independent Variable Employee Risk-taking Propensity ________________________ 18 Mediating Variable Employee Voice Quantity________________________________ 19 Moderating Variable Employee Working Memory Capacity _____________________ 19 Procedure ______________________________________________________________ 19 Analyses and Predictions __________________________________________________ 20 Figure 1. Measurement Model ______________________________________________ 21 Results ___________________________________________________________________ 21 Exploratory Factor Analysis and Reliabilities __________________________________ 21 Dependent Variable Employee Voice Quality ________________________________ 22 Independent Variable Employee Risk-taking Propensity ________________________ 23 Mediating Variable Employee Voice Quantity________________________________ 24 Correlations _____________________________________________________________ 25 Table 1. Descriptives, Correlations, and Cronbach’s Alphas ______________________ 27 Testing the Mediation Model _______________________________________________ 28 Table 2. Process Analysis __________________________________________________ 29 Table 3. Process Analysis – Mediation ________________________________________ 29 Testing the Moderated Mediation Model ______________________________________ 30 Table 4. Process Analysis – Moderated Mediation ______________________________ 31 Table 5. Process Analysis – Conditional Indirect Effects at Values of the Moderator ___ 31 Figure 2. Effects Model ___________________________________________________ 32 Discussion ________________________________________________________________ 32 Summary of Study Results _________________________________________________ 32 Alternative Explanations ___________________________________________________ 34 Methodological Points of Critique and Positive Points ___________________________ 36 Suggestions for Future Research ____________________________________________ 37 Interpretations and Contributions ____________________________________________ 39 Concluding Thoughts _____________________________________________________ 41 References ________________________________________________________________ 42 Appendix A – Employee Voice Quality Items ____________________________________ 48 Appendix B – Employee Voice Quantity Items ___________________________________ 50 Appendix C – Employee Risk-taking Propensity Items _____________________________ 51
4 Abstract
Where previous studies on employee voice behavior focus on how often (voice quantity) employees communicate ideas, suggestions, problems, and opinions, our study focuses on the quality (voice quality) of these ideas, suggestions, problems, and opinions. More specifically, we study how employees’ risk-taking tendency and cognitive capacity influence the originality of voiced messages. Since voice behavior is seen as risky decision-making, we hypothesize that employees who tend to take more risks voice more often. Furthermore, we propose that these risk-taking employees who voice more often improve the quality of their output by practicing voice. Last, we hypothesize that these risk-taking employees indeed achieve higher voice quality through voicing more often, but only when they also posses the cognitive capacity to focus on and process relevant information. One-hundred-and-seventy-two employees and eighty-four supervisors participated in our study, which included an online survey and a visuo-spatial working memory capacity task. The results show that employees who tend to take more adventurous risks, such as recreational, social, or career risks, indeed communicate ideas, suggestions, problems, and opinions more often, and achieve more original output, independent of their cognitive capacity. We conclude that organizations could benefit from hiring adventurous risk-taking employees and offering enough space to practice and achieve original voice.
5 Introduction
As a supervisor you deal with different types of employees: lazy employees, chatty employees, intelligent employees, finish-the-job-and-go-home employees, proactive employees, and employees who like to take risks in their career. But what type of employee would you prefer to work in your company? Employees who come in, finish their job tasks and go home without trying to change things within the organization or employees who take risks and go the extra mile by coming up with and communicating original ideas to improve the organization?
The latter type of employee is one that engages in voice behavior, whereby employees take the initiative to communicate ideas, suggestions, problems, and opinions to their supervisors (Morrison, 2014, 2011; Detert & Burris, 2007; Van Dyne & LePine 1998). The phenomenon of employee voice behavior is studied extensively (Morrison, 2014, 2011), and is shown to contribute to organizational success in the form of financial performance (MacKenzie, Podsakoff, & Podsakoff, 2011) and unit effectiveness (Detert, Burris, Harrison, & Martin, 2013). Although this indicates that it is indeed important for organizations to have employees who communicate innovative ideas, extant literature focuses only on how often employees communicate their ideas, suggestions, problems, and opinions (voice quantity), not taking into account how useful and original and thus potentially innovative, these ideas, suggestions, problems, and opinions are (voice quality, e.g. Wolsink, 2013). Wolsink (2013) suggests that the quality of voice is of equal or even of more value for organizational success than the quantity of voice, since organizations are likely to improve more when employees think about the problems, opinions, ideas, and suggestions they have and communicate the ones that are useful and original, and thus innovative.
Since this quality subject is still relatively unexplored, it seems important to know what factors might influence the quality of employee voice. One of the factors that always
6 come up in voice theory, but is rarely empirically investigated in relation to voice (and never in relation to voice quality) is risk. Although voice can be beneficial for the organization, it could also harm the employee who voices. For example, voice can entail social risk, because supervisors do not always respond well to employee voice (Detert & Burris, 2007). Furthermore, extant literature assumes that employees make conscious risk calculations in their decision to voice or not (Morrison, 2011). Therefore, one of the factors that are likely to affect employee voice is an employee’s evaluation of risk. Employees take into account the risk of communicating ideas, suggestions, problems, and opinions to their supervisors, and they become reluctant to voice when they perceive risks as high (Lui, Zhu, & Yang, 2010; Detert & Burris, 2007; Milliken, Morrison, & Hewlin, 2003). However, these studies focus primarily on the perceptions of risk, while actual risk-taking tendencies of employees are not yet studied within the voice literature. This is important, since employees with a propensity to take risks might make different decisions regarding the communication of ideas, suggestions, problems and opinions since they might ignore the high risk of voicing or might perceive it differently. It seems important to also take into account the actual tendencies of employees regarding their taking, because which employee will voice more? The impulsive risk-taker or the adventurous risk-risk-taker? With the intention to study a new topic within the voice literature, we suggest that employees with high risk-taking propensity are likely to voice more frequently (i.e., voice quantity). Furthermore, their voice increases in quality over time, because they practice more and receive more feedback than employees with low risk-taking propensity.
Another factor that is likely to influence the quality of employee voice is the ability to control and focus attention (i.e. working memory capacity). Working memory capacity is referred to as the system that is responsible for keeping information active for complex cognitive activities, and is shown to benefit creative thinking (De Dreu, Nijstad, Baas,
7 Wolsink & Roskes, 2012). Therefore, it is likely that an employee’s ability to control attention and to focus on information that is relevant determines the usefulness and originality, both facets of creativity (Runco & Jaeger, 2012), of communicated ideas, suggestions, problems, and opinions. Hence, it seems possible that employees who have a high tendency towards taking risks also communicate more and thus communicate better ideas, suggestions, problems, and opinions when they also have the ability to focus on and process the relevant information. In general, what would risk-taking be without thinking things through? Recklessness, and this not likely to be appreciated on the work floor. Thus, working memory capacity could serve as a guide to direct risk-taking behavior in the right direction.
Employee voice behavior is a well-studied subject, however, most studies only address how often employees communicate ideas and suggestions, leaving out the originality and usefulness (i.e. quality) of these ideas, suggestions, problems, and opinions. Moreover, the existing literature only theorizes about the influence of risk perceptions on employee voice behavior, and does not address how employees’ actual risk-taking tendencies influence the communication of high quality ideas, suggestions, problems, and opinions. Furthermore, extant literature shows that there is a link between working memory capacity and creativity, and only two studies (Wolsink, Den Hartog, Belschak, & Sligte, 2015; Chiaburu, Marinova, & Van Dyne, 2008) address the link between cognition and employee voice behavior. Moreover, how cognitive ability and risk interact in predicting voice behavior is interesting because communicating more often doesn’t necessarily have to result in high quality ideas, suggestions, problem, and opinions. To get from frequent voice to quality voice, high risk-taking employees might depend on their cognitive capacity for information processing. The purpose of this study is to address these research gaps and contribute to the literature by (1) differentiating between the two components of employee voice behavior, namely quantity
8 and quality, and (2) adding knowledge about the influences of risk to the existing literature on employee voice. The following research question will guide this study in fulfilling this purpose: “Do risk-takers learn how to voice original ideas through practice? And does cognitive capacity enhance that process?”
Theoretical Framework The Challenging Nature of Voice
Employee voice behavior is often defined as the communication of innovative ideas and suggestions, opinions, concerns, problems, and possibilities for improvement within the organization to someone who is able to take the appropriate action (Morrison, 2014, 2011; Tangirala & Ramanujam, 2008; Detert & Burris, 2007; Van Dyne & LePine, 1998).
Although the above definition states that employee voice behavior is targeted towards people higher up in the organization, it can also be directed towards colleagues. Voicing towards colleagues is referred to as speaking out, whereas employee voice behavior directed to supervisors is called speaking up (Morrison, 2014; Lui, Zhu, & Yang, 2010; Detert & Burris, 2007). Thus, employee voice can be distinguished based on target group. Another distinction of employee voice can be based on the content of voice. Employee voice behavior can be divided in three types, namely opinion focused voice, problem focused voice, and suggestion focused voice (Morrison, 2011). The latter, suggestion focused voice, covers the communication of ideas and suggestions for improvements within the organization. Problem focused voice involves the communication of concerns about the organization. Opinion focused voice concerns the communication of opinions that differ from those of others within the organization. All three types of voice are included in the definition of employee voice behavior as given above, and we will use this definition throughout the paper.
9 Besides having different targets and content types, employee voice behavior is always an extra-role behavior (Van Dyne & LePine, 1998) that is both challenging and promotive (Chiaburu, Marinova, & Van Dyne, 2008; Van Dyne & LePine, 1998). Employees who voice do this voluntarily and not because it’s part of their job requirements. They have the intention to challenge and change the status quo within the organization, but do so with constructive intend. Employees who communicate ideas, suggestions, problems, and opinions are pro-actively encouraging things to happen and change (Chiaburu, Marinova, & Van Dyne, 2008; Van Dyne & LePine, 1998). This challenging description of employee voice behavior is also why voice is seen as a risky behavior (Lui, Zhu, & Yang, 2010; Detert & Burris, 2007; LePine & Van Dyne, 1998), since others in the organization do not always appreciate change (Samuelson & Zeckhauser, 1988). Perceptions of risk are thus likely to influence employees’ decisions to voice.
Moreover, extant literature assumes that employees consciously calculate the risks in these decisions to voice or not, by weighing the costs, benefits, and potential success of voice (Morrison, 2014, 2011). Employees are more likely to engage in voice behavior when they perceive the benefits and potential success of voicing as high and the risks (safety) as low (high) (Morrison, 2014). This is why, for example, self-efficacy (confidence) predicts the frequency of voice. Although employees might make these kinds of calculations and perceive voice behavior as risky with potential undesirable outcomes, the relationship between their actual risk-taking tendencies in other parts of life and voice behavior is unknown. It seems important to know how actual risk-taking tendencies influence employee voice rather than just risk perceptions, since two employees might react differently to the same risk perceptions. Employees with a tendency to take risks might react differently to a risky voice opportunity than employees who have a tendency to avoid risks.
10 Risky Situations or Risky People?
Risk, as a characteristic of a decision, can be defined as the extent to which the outcome of that decision is uncertain, as the results of that decision can be either desirable or disappointing (Sitkin & Pablo, 1992). Desirable outcomes will decrease (increase) the risk (safety) of voicing, whereas disappointing outcomes will increase (decrease) the risk (safety) of voice in the calculations employees make before either voicing or remaining silent.
Risky decision-making is a broadly studied subject in several disciplines, for example, economics, biology, and (cognitive) psychology (Mishra, 2014). The current literature holds two views concerning the factors influencing risky decision-making, namely individual differences and situational factors. Situational factors, such as low income, are immediate causes of risky behavior and can vary within people regarding the situation they are in, whereas individual differences result from development and genetics, and over the lifetime, develop into a stable tendency towards either risky or safe behavior (Mishra, 2014). The latter is described by risk propensity theory, whereas the former is described by risk-sensitivity theory.
Risk propensity is defined as the tendency to either take or avoid risks (Sitkin & Weingart, 1995; Sitkin & Pablo, 1992). Therefore, risk-taking propensity can be defined as the tendency to take risks. High risk-taking propensity is shown to result in more risky decision-making behavior through lowering risk perception (Sitkin & Weingart, 1995). In other words, a decision-maker with higher risk-taking tendencies perceives the risk of that particular decision differently, which results in a different behavior compared to one with lower risk-taking tendencies. For example, someone with high risk-taking propensity might gamble more frequently than someone with low risk-taking propensity, because he or she perceives the risk of losing money differently. Thus, it is interesting to investigate whether the tendency to take risks influences employees’ decision making regarding voice.
11 Employee Voice as Risky Decision-making Behavior
Employee voice behavior is seen as a risky behavior (Lui, Zhu, & Yang, 2010; Detert & Burris, 2007; LePine & Van Dyne, 1998), because outcomes are uncertain. Therefore, the decision whether to voice or not can be seen as a risky decision. A risky decision has one of the following characteristics: (1) goals are difficult to achieve, (2) expected outcomes are uncertain, or (3) potential outcomes include extreme consequences (Sitkin & Weingart, 1995). Looking at voice behavior, we can argue that all three characteristics fit. First, the goal of voice is change, which can be difficult to achieve in organizations (Schneider, Brief, & Guzzo, 1996). Second, the outcomes of voice are uncertain, since supervisors might react either positively or negatively to critique in the form of ideas and suggestions, problems, opinions, and points for improvement (Detert & Burris, 2007). Third, potential outcomes of voice behavior can be extreme, such as a negative public image (Milliken, Morrison, & Hewlin, 2003) or damaged relationships with others (Adler & Kwon, 2002). Moreover, voice behavior is also risky from a resource-perspective: voiced ideas either get implemented or not, so why put in the effort if nothing changes?
The above characteristics of risky decision-making regarding employee voice behavior can make employees reluctant to engage in voice behavior. Employees can perceive the risk of voicing as high when they believe that the achieving the change is difficult, their supervisors might react negatively, and relationships with others become damaged. However, it is possible that employees with high risk-taking propensity perceive this risk as lower or might even ignore it, and therefore engage in risky voice behavior more often than employees with low risk-taking propensity. Another example of how risk-taking propensity might lead to risky behavior on the work floor is that an employee who tends to take risks will go and ask the supervisor for a promotion or for an opportunity to work a few months abroad, whereas an employee who tends to avoid risks might just wait for such opportunity to fall
12 onto his or her desk. These are examples of career risks, but it would also be interesting to look at the effects other types of risks have on employee voice behavior. Therefore, we hypothesize that:
H1: Employees with high risk-taking propensity are likely to engage in risky behaviors on the work floor, and therefore are likely to communicate ideas and suggestions more often compared to employees with low risk-taking propensity.
The Importance of Voice Quality for Organizations
Although the above definition of employee voice behavior captures the quality of voice as innovativeness, studies that actually measure the quality of voice remain scarce because the voice measures are designed to measure how often employees engage in voice behavior and not how good their voice is (Wolsink, 2013). An example of such a study is that of Van Dyne and LePine (1998), in which they used a survey including the following item: “This particular employee develops and makes recommendations concerning issues that affect this organization.” This item (and the other items alike) measures the quantity of employee voice, but does not include its quality. However, it seems that the quality aspect of employee voice is more important for organizations than the quantity aspect, since being innovative is important for organizational survival (Cefis & Marsili, 2005; Christensen, 1997).
Wolsink (2013) defines voice quality as the communication of ideas and suggestions that are useful and original. This definition is based on facets of creativity: ideas are seen as creative when they are both original and useful (Runco & Jaeger, 2012). Amabile (1988) argues that creativity (idea generation) is important for innovation (idea implementation) within organizations. An important link in this idea generation-implementation process is the communication of creative, and thus original and useful ideas, which is referred to as voice
13 quality. Although voice quality and creativity are similar in content, they are distinct in process. Creativity can be seen as the idea generation process, whereas voice quality can be seen as the communication of these ideas, and innovation as the implementation of these ideas. Employees might generate original and useful ideas, but these cannot be implemented correctly if they are not communicated (i.e. voice quality). Thus, it is likely that voice quality is the step needed to get from generating original and useful ideas to implementing these ideas. So, following Amabile’s (1988) line of thought, we can argue that this research brings us one step closer to innovation, by identifying an important link between creativity and innovation and showing what influences it.
How Risk-taking Propensity Might Lead to High Quality Voice Through Practice
Employee voice quantity is shown to affect voice quality (Wolsink et al., 2015; Wolsink, 2013), indicating that employees who engage in voice behavior often also reach higher voice quality compared to employees who voice rarely. A reason for this might be that employees who voice more frequently receive more feedback, which enhances individual learning (London, 2003) and might enhance voice quality as well. Therefore we can argue that employees with high risk-taking propensity are likely to achieve high voice quality in the long run through practice by engaging in voice more often than employees with low risk-taking propensity.
Moreover, existing literature argues that risk-taking is fundamental for innovation (Ford and Gioia, 1995), and that it distinguishes radical innovations from incremental ones (Ekvall, 1996). Moreover, risk-taking also influences creativity, more specifically, risk-taking is seen as a personality prerequisite for creative thinking (Sternberg & Lubart, 1992, 1991). Thus we can argue that risk-taking propensity not only affects the quality of employee voice
14 through increasing voice quantity, but that it also might cause employees to think out of the box. We hypothesize that:
H2: Employees with high risk-taking propensity are likely to voice more often, practicing voice behavior more frequently, and therefore are more likely to communicate ideas and suggestions that are original and useful compared to employees with low risk-taking propensity. We thus propose that a positive relationship between risk-taking propensity and voice quality is mediated by voice quantity.
How Cognitive Ability Might Foster Voice Quality
Attention control is the ability to control attention and to focus on information that is relevant while doing different things (Kane & Engle, 2003, 2002), and is seen as a necessity for creative thinking (Dietrich, 2004). Essential for attention control is working memory capacity (Kane & Engle, 2003, 2002), which is referred to as the system responsible for keeping information active for difficult cognitive activities (De Dreu et al., 2012). Working memory capacity helps people to manage the information available to them (Oberauer & Bialkova, 2009) in a controlled and conscious way (Evans, 2008), which helps creative insight (De Dreu et al., 2012) and performance of learning behavior (Riding, Grimley, Dahraei, & Banner, 2003).
The working memory system consists of one overarching system responsible for controlling attention, the central executive, and two subsystems responsible for temporary storage of information, the visuo-spatial sketchpad and phonological loop (Baddeley, 2012; Baddeley & Hitch, 1974). The central executive system is responsible for controlling and integrating information from both subsystems (Riding et al., 2003). The phonological loop (verbal working memory) is responsible for speech-based information, whereas the
visuo-15 spatial sketch (visual working memory) is responsible for holding information on both locations and visual appearances (such as color or shape) of objects (Baddeley, 1998). Visual working memory is able to store integrated object features, meaning that it has a large capacity for holding a lot of individual features of a small amount of objects (Luck & Vogel, 1997). In other words, employees with high visual working memory capacity will be able to store and process different pieces of information regarding one specific problem, for example a broken machine at work. This will allow them to come up with better, more creative, solutions than employees with low visual working memory capacity, who are not able to integrate the available information.
Furthermore, De Dreu, Baas and Nijstad (2008) argue that creativity can be achieved through two ways, namely cognitive persistence and cognitive flexibility. The latter is associated with reduced attention control, by means of which one can switch to different approaches and perspectives, whereas the former is associated with controlled attention focus and systematic thinking (Nijstad et al., 2010). Since working memory capacity is central to one’s ability to (systematically) control and focus attention, it benefits creative thinking through the cognitive persistence path. Someone with high working memory capacity will be better able to block distractions, focus on and process relevant information in a persistent manner, and is therefore more likely to solve complex problems or come up with innovative solutions than someone with low working memory capacity. Since working memory capacity fosters the generation of creative ideas and suggestions, it is also likely to increase their voice quality through this generation process. Furthermore, Wolsink and colleagues (2015) found that working memory capacity also helps voice quality because it allows for selection of the best ideas for voice. In their study, they distracted participants, which then influenced their selection of ideas to communicate. Participants with higher working memory capacity were
16 better able to block the distractions and select the best ideas for communication than participants with lower working memory capacity.
Risk-taking is Good, But Only When Working Memory Capacity is High
As mentioned before, voice is seen as risky decision making, whereby employees make calculated decisions. Moreover, cognitive skills are argued to influence one’s risk-taking behavior such that people high in cognitive ability are more willing to take calculated risks (Burks, Carpenter, Goette, & Rustichini, 2009). However, cognitive skills can be impeded by, for example, a state of poverty (Mani, Mullainathan, Shafir, & Zhao, 2013). In their study, Mani and colleagues (2013) found that poverty-related concerns put a load on one’s cognitive capacity, leaving less “space” in one’s mind to think about other things or process information. Thus, employees who feel that they are in a state of poverty, or undesired state, might have less cognitive capacity left for processing available information because of concerns about this state. They have too much going on in their heads (i.e., a high load on their working memory capacity), which can result in bad decision making due to safety or risk miscalculations. Thus, employees high in working memory capacity might make better risk calculations regarding the outcome of their voice behavior. In other words, they might be better able to calculate whether communicating an idea would, for example, harm the relationship with their supervisors or not.
Furthermore, taking into account that working memory capacity is likely to benefit the creative thinking and learning of employees, we can argue that risk-taking propensity might positively influence the quality of voice through the quantity of voice only when employees also have high working memory capacity. Employees with high risk-taking propensity and low working memory capacity can voice as often as those with high working memory capacity, but might not be able to focus their attention and process the information
17 relevant to, for example, a certain problem or process in an effective way. This would then restrict them from achieving the same quality ideas, suggestions, problems, and opinions through practice. For example, employees A and B both have more or less the same risk-taking tendencies when it comes to their career or social life, but they differ in their cognitive ability, such that employee A is better able to block distractions, focus on and process available information, and learn from similar problems in the past (for example, through processing feedback) than employee B. They both work on project X and encounter the same problem. Although both have the same information regarding this problem available to them, employee A will be better able to process this information into a more original solution to the problem than employee B. Therefore, we hypothesize that:
H3: Employees with high risk-taking propensity achieve high(er) voice quality through voice quantity when they also posses high working memory capacity. Employees with low working memory capacity might not achieve the same quality of voice, even though they also engage in more (risky) voice behavior.
Method Design and Sample
The study had a cross-sectional design with both an online survey and a visual working memory task on an iPad Mini with Retina display. Data was collected in triads, each consisting of one supervisor and two employees. Participants were sampled through personal contacts of bachelor and master students of the Amsterdam Business School.
The sample consists of Dutch-speaking employees (N = 172) and supervisors (N = 84) residing in the Netherlands. Data was collected on team basis, consisting of one supervisor
18 and two employees who work together on a daily basis. Two employees were removed from the sample due to either missing or unreliable data.
This resulted in a final sample of 170 employees (47.3% female). This was a good representation of the Dutch workforce (48% female: CBS, 2015, February 26). The average age of employees is 32.38 (SD = 12.72), while the average tenure is 6.01 years (SD = 8.67). Employees’ working memory capacity score was 6.27 on average (SD = 1.03). Of the supervisors, 67.9% is male, the average age is 39.92 (SD = 12.78), and the average tenure is 9.77 years (SD = 9.68). Most of the employees worked in a private company (32.5%), followed by the retail sector (18.1%) and the non-profit sector (13.3%).
Measurements
Dependent Variable Employee Voice Quality
Employee Voice Quality was measured with both supervisor and colleague ratings using a 20-item scale developed by Wolsink, Den Hartog, Belschak and Sligte (2015). The scale covers both usefulness and originality aspects of voice quality equally. Example items are: ‘When my employee/colleague comes up with an idea, it is original’ (originality) and ‘When my employee/colleague gives suggestions, it is useful to others’ (usefulness). Scales range from (1) not at all applicable to (7) entirely applicable. A high score means that the employee communicates ideas and suggestions that are original and useful.
Independent Variable Employee Risk-taking Propensity
Employee Risk-taking Propensity was measured at employee level with a 12-item scale developed by Nicholson and colleagues (2005). The scale covers six different domains, namely recreational, career, social, health, financial, and safety, and includes present and past risk-taking behavior. The latter three domains seem to cover a more self-destructive type of
19 risk-taking, whereas the first three domains cover a more adventurous type of risk-taking. An example item is: ‘I take career risks’. Scales range from (1) never to (5) very often. A high score means that the employee has a high tendency to take risks.
Mediating Variable Employee Voice Quantity
Employee Voice Quantity was measured on supervisor and colleague level with a 11-item scale developed by Wolsink, Den Hartog, Belschak and Sligte (2015). An example item is: ‘My employee/colleague provides solutions for problems within the organization’. Scales range from (1) never to (7) very often. A high score indicates that the employee communicates ideas and suggestions frequently.
Moderating Variable Employee Working Memory Capacity
Employee Working Memory Capacity was measured at employee level with a visual working memory task conducted on an Apple iPad Mini with Retina Display developed by Wolsink, Den Hartog, Belschak and Sligte (2015). The task consisted of eight trials where employees saw footballs and tennis rackets in a five by five grid, and they need to remember the places of the footballs and to remember which of the rackets turned. Each trial increases in difficulty, for example by including distractors, and runs until two mistakes are made. A high score on this task means that the employee has a high working memory capacity, indicating that he or she is capable of controlling and focusing attention.
Procedure
Bachelor and master students of the Amsterdam Business School contacted one of the team members personally, who then arranged a meeting with the whole team. On the day of the appointment, one or two researchers went to the workplace of the participants where they
20 administered the survey and visual working memory task at the location. The researcher(s) asked for a closed room where employees could take the visual working memory task to ensure that employees were fully focused and were not disturbed by others during the task. The online surveys were taken on the computers of the participants or the researchers. The researcher(s) shortly explained the process of the online survey and iPad task to the whole team. While the supervisor and one of the employees started with the online surveys, the other employee started the iPad task after receiving instructions on how to hold the iPad. This is important since holding the iPad differently might have decreased reaction time.
Analyses and Predictions
The data was analyzed with a regression-based path analysis by using PROCESS (Hayes, 2013), an existing tool for testing mediation and moderation models. First, we tested the mediation effect of Employee Voice Quantity in the relationship between Employee Risk-taking Propensity and Employee Voice Quality (H2). We predicted a direct positive effect of employees’ risk-taking propensity on their voice quality. This means that we expected to find that employees with high risk-taking propensity would also communicate more useful and original ideas. Furthermore, we predicted a stronger indirect positive effect of employees’ risk-taking propensity on their voice quality, mediated by their voice quantity. In other words, we expected to find that employees with high risk-taking propensity would voice more often, and therefore communicate more useful and original ideas.
Next, we tested the moderated mediation effect of Employee Working Memory Capacity on the mediation model (H3). We predicted that employees’ working memory capacity moderates the relationship between their risk-taking propensity, voice quantity and voice quality. In other words, we expected that employees with high risk-taking propensity would voice more often and achieve high voice quality, but only when their working memory
21 capacity is also high. However, employees with high risk-taking propensity and with low working memory capacity would not be able to achieve the same level of voice quality, even if they voice equally frequent.
We tested these models by using the bootstrapping method with 5000 interactions. The program tests 50% of the data sample, repeating the process for 5000 times with a random 50% of the sample.
Figure 1. Measurement Model
Results Exploratory Factor Analysis and Reliabilities
An exploratory factor analysis with Direct Oblimin rotation was conducted on three variables, namely employee voice quality, risk-taking propensity, and voice quantity. This
11 7-point Likert items, rated by supervisors and colleagues, from Wolsink, Den Hartog, Belschak and
Sligte (2015) Employee Voice Quantity Employee Risk-taking Propensity
12 5-point Likert items, rated by employees, from Nicholson, Soane, Fenton‐ O'Creevy & Willman (2005)
Employee Voice Quality
20 7-point Likert items, rated by supervisors and colleagues, from Wolsink, Den Hartog, Belschak and
Sligte (2015)
Employee Working Memory
Capacity
Visuo-spatial task o
trial types, run
mistakes are made, from Wolsink, Den Hartog, Belschak and Sligte (2015)
22 rotation type was chosen, because different factors within each variable are expected to be related. Factors are based on Eigenvalue (> 1).
Moreover, we conducted a reliability analysis to see whether the measurements mentioned in the method section can be used for our regression analysis. All alphas can be found in Table 1.
Dependent Variable Employee Voice Quality
Based on creativity theory and Wolsink and colleagues (2015), we expected employee voice quality to divide into two categories (for items see Appendix A). This was supported by the results of the factor analyses, which show that voice quality can be distinguished into two factors, originality and usefulness. Because this variable was measured from two points of view, namely supervisor and colleague, we conducted two separate exploratory factor analyses. Variances are given below and are from supervisor and colleague point of view respectively.
From both points of view, the analysis indeed shows that there are two factors within the voice quality construct, one measuring the originality (8.508% and 10.675% of variance) and one measuring the usefulness (64.047% and 54.876% of variance) of employee voice, together accounting for 72.555% and 65.551% of the variance.
Two supervisor- and colleague-rated originality items (items 7 and 9) load on both factors, with little difference between the usefulness and originality factor, so these items were deleted. Item 7 captures how “refreshing the advice is”, whereas item 9 measures if “advice gives new insights”. Furthermore, one of the supervisor-rated items that should measure originality (item 5) loads more on the usefulness factor, so this item was also deleted. This item measures if “the solutions for problems are original”. Moreover, one of the
23 colleague-rated items that is supposed to measure originality (item 4) loads more on the other factor, so this item was also deleted. This item measures if a “suggestion is often renewing”.
This resulted in a clear separation of the originality (9.644% and 12.124% of variance) and usefulness (62.527% and 53.596% of variance) factors, accounting for 72.171% and 65.719% of the variance together. The Kaiser-Meyer-Olkin measure of sampling adequacy is 0.952 for supervisor ratings and 0.916 for colleague ratings. Concluding, we chose to divide voice quality into the originality and usefulness factors that came forward from this analysis. Cronbach’s alphas for both factors are good, 0.897 and 0.936 respectively.
Independent Variable Employee Risk-taking Propensity
This variable initially intended to measure employee’s risk-taking propensity over six different dimensions (for items see Appendix C). Because this variable measured risk-taking propensity for two different periods, namely past and present, we conducted two separate factor analyses, one for past risk-taking propensity and one for present risk-taking propensity. Given variances are for past and present risk-taking propensity respectively.
The factor analyses for past and present risk-taking propensity both show that there are two factors on which the items load, accounting for 56.190% and 57.553% of the variance. One factor seems to measure the more dark side of risk-taking (38.710% and 40.163% of variance), whereby taken risks are more self-destructive. This factor includes health, financial, and safety risks (items 2, 4, and 5). The other factor seems to reflect a more bright side of risk-taking (17.480% and 17.390% of variance), whereby taken risks are more adventurous. This factor includes recreational, career, and social risks (items 1, 3, and 6). The Kaiser-Meyer-Olkin measure of sampling adequacy is 0.738 for past ratings and 0.725 for present ratings.
24 We will conduct further analysis with the bright and dark side of risk-taking propensity, and because past and present risk-taking correlate highly for both factors, r (166) = 0.857, p < 0.01 for bright side and r (166) = 0.889, p < 0.01 for dark side risk-taking, we will take past and present risk-taking together. Cronbach’s alphas for both dark and bright side risk-taking propensity are good, 0.824 and 0.793 respectively.
Mediating Variable Employee Voice Quantity
Based on voice theory, we expected employee voice quantity to fall into three categories, namely suggestion, problem, and opinion focused voice (for items see Appendix B). We conducted two separate factor analyses because this variable was measured from both supervisor and colleague point of view.
From both points of view, the results do not show such a tripartition. The factor analysis results in one factor for supervisor ratings, accounting for 46.084% of the variance. Looking further to the communalities, the results show that two items (items 6 and 10) have communalities lower than 0.4. Deleting these two items results in an explained variance of 50.994%. For colleague ratings the factor analysis shows two factors, accounting for 63.435% of the variance. One factor includes the suggestion and most of the opinion focused voice items (items 1, 2, 3, 4, 5, 6, and 9), whereas the other factor includes most of the problem focused items (items 7, 8, 10, and 11). The Kaiser-Meyer-Olkin measure of sampling adequacy is 0.913 for supervisor ratings and 0.886 for colleague ratings.
Looking at the results of the factor analysis from both points of view, we conclude that there are no clear factors present, so we decide to analyze further with the employee voice quantity variable as a whole. The Cronbach’s alpha for voice quantity is good, 0.913.
25 Correlations
We conducted a correlation analysis, of which the results can also be found in Table 1. Unexpectedly, there is no correlation between employee voice quantity and employee risk-taking propensity, either dark (r (166) = 0.055, ns) or bright side (r (166) = 0.123, ns). This means that employees who tend to take more risks, may it be self-destructive or adventurous in nature, do not also communicate more ideas, suggestions, problems, or opinions than employees who take lower risks.
Moreover, there is also no correlation between employees’ dark side risk-taking propensity and voice quality (r (166) = 0.018, ns), neither with the originality (r (166) = 0.103, ns) or usefulness (r (166) = -0.072, ns) aspects of quality. However, there is a correlation between employees’ bright side risk-taking propensity and voice quality (r (166) = 0.158, p < 0.05), and with its originality aspect (r (166) = 0.197, p < 0.05), whereas there is no correlation with its usefulness aspect (r (166) = 0.094, ns). This partially confirms our expectations, suggesting that employees who take more adventurous risks also communicate more original ideas, suggestions, problems, and opinions. However, employees who take more self-destructive risks do not necessarily communicate ideas, suggestions, problems, or opinions that are original and/or useful.
Furthermore, the results confirm the expected relationship between employee voice quantity and voice quality (r (170) = 0.622, p < 0.01), for both originality (r (170) = 0.539, p < 0.01) and usefulness (r (170) = 0.609, p < 0.01). This indicates that employees who voice more often also communicate more original and/or useful ideas, suggestions, problems, or opinions than employees who voice less often.
Moreover, the results show the expected correlation between employees’ working memory capacity and the originality aspect of voice quality (r (170) = 0.180, p < 0.05), whilst there is no correlation with voice quality (r (170) = 0.135, ns) or with the usefulness aspect (r
26 (170) = 0.068, ns). This means that employees’ working memory capacity is only related to the originality of the communicated ideas, suggestions, problems, and opinions and not to the usefulness of voice. Employees with higher working memory capacity thus communicate ideas that are more original than employees with low working memory capacity.
The results also show a surprising correlation between employees working memory capacity and their risk-taking propensity, which is stronger for dark side risks (r (166) = 0.273, p < 0.01) than bright side risks (r (166) = 0.174, p < 0.05). Thus, employees with higher working memory capacity also have higher self-destructive and adventurous taking propensity. Rational thinking suggests that age plays a role herein, since both risk-taking propensity (Deakin, Aitken, Robbins, & Sahakian, 2004) and working memory capacity (Dobbs & Rule, 1989) can change with age.
Since working memory capacity, bright side risk-taking and age correlate so highly with each other, we decided to control for age in further analysis.
27 Table 1. Descriptives, Correlations, and Cronbach’s Alphas on Diagonal (N = 170).
Variables M SD 1 2 3 4 5 6 7 8
1. Voice Quality (DV) 4.446 0.870 (0.951)
2. Voice Quality – Usefulness (DV) 4.739 0.936 0.922** (0.939)
3. Voice Quality – Originality (DV) 4.153 0.950 0.924** 0.703** (0.897)
4. Bright Side Risk-taking (IV) 2.012 0.770 0.158* 0.094 0.197* (0.793)
5. Dark Side Risk-taking (IV) 2.309 0.881 0.018 -0.072 0.103 0.420** (0.824)
6. Voice Quantity (MeV) 4.469 0.854 0.622** 0.609** 0.539** 0.123 0.055 (0.913)
7. Working Memory Capacity (MoV) 6.270 1.033 0.135 0.068 0.180* 0.174* 0.273** 0.037 (0.811)
8. Age (CV) 32.38 12.72 -0.081 -0.007 -0.140 -0.311** -0.391** 0.091 -0.489** -
28 Testing the Mediation Model
We used Process regression analysis in order to examine whether the proposed mediating effect of employees’ voice quantity in the relationship between risk-taking propensity and voice quality is present. We conducted the analysis with the bright side risk-taking propensity of employees and the originality aspect of employee voice, since only these factors of risk-taking propensity and voice quality are related in our sample.
In alignment with our first hypothesis, the results (Table 2, Model 2) show a direct effect of employees’ (bright side) risk-taking propensity on their voice quantity (β = 0.189, t = 2.112, p < 0.05, R2 = 0.037), when controlled for age. This indicates that employees who take more adventurous risks, such as career, social, or recreational risks, also communicate more ideas, suggestions, problems, or opinions than employees who do not take such risks.
Furthermore, the results (Table 2, Model 1) show a positive effect of employees’ bright side risk-taking propensity on the originality of their voice (β= 0.216, t = 2.573, p < 0.05, R2 = 0.043). This is also in line with the correlation results and means that employees who take more adventurous risks also communicate ideas, suggestions, problems, or opinions that are more original than employees who do not take such risks.
Moreover, the results show a full mediating effect (Table 3, Model 3) of employees’ voice quantity in the relationship between their risk-taking propensity and the originality of their voice, with an indirect effect of 0.113 (CI: 0.000; 0.258), while there is no direct effect (0.103) of employees’ bright side risk-taking propensity on the originality of their voice anymore (CI: -0.064; 0.271). The total effect of 0.216 is also present (CI: 0.020; 0.412). This suggests that employees with higher bright side risk-taking propensity reach more original voice because they communicate ideas, suggestions, problems, or opinions more often.
29 Table 2. Process Analysis
Model 1 Model 2
Dependent Variable Voice Quality – Originality Voice Quantity
Coefficient (β) SE t Coefficient (β) SE t
Constant 3.885** 0.203 18.062 3.751** 0.296 12.667
Bright Side Risk-taking 0.216* 0.094 2.573 0.189* 0.089 2.112
Age -0.005 0.006 -0.853 0.010 0.005 1.903
R2 0.043 0.037
Note. N = 166. * p < .05. ** p < .001.
Table 3. Process Analysis – Mediation
Model 3
Dependent Variable Voice Quality – Originality
Effect SE LLCI ULCI t
Indirect Effect 0.113* 0.065 0.000 0.258
Direct Effect 0.103 0.085 -0.064 0.271 1.216
Total Effect 0.216* 0.099 0.020 0.412 2.178
30 Testing the Moderated Mediation Model
Furthermore, we wanted to look if a moderator could influence the proposed model, especially since the correlation analysis shows that working memory capacity, voice quality-originality and bright side risk-taking are related (see Table 1). For our dependent variable we again only looked at the originality aspect of employee voice quality, and for our independent variable we again looked at the bright side of risk-taking propensity.
Unfortunately, we did not find a moderation effect when controlled for age (Table 4, Model 4) of employee working memory capacity on the above proposed mediation model (β = -0.020, t = -0.314, ns, R2 = 0.327). This indicates that the presence of mediation in our sample is not contingent on employees’ working memory capacity. Thus, employees who take more adventurous risks will voice more original ideas, suggestions, problems, or opinion through communicating more, regardless of their working memory capacity.
However, more in-depth analysis revealed that the interaction effect seems present for high employee working memory values (see Table 5). The results demonstrate that for employee working memory capacity values of 6.283 (M, 0.111, CI: 0.001; 0.255) and 7.320 (+ 1 SD, 0.107, CI: 0.005; 0.277) there is indeed an indirect effect of employees’ bright side risk-taking propensity on the originality of their voice, while the direct effect (0.105) is not present (CI: -0.064; 0.275, Table 4). However, the mediating effect is weaker for those with higher working memory capacity compared to those with lower working memory capacity. This strengthens the indication that employees’ working memory capacity hardly influences the mediation model. However, it indicates that it might be worthwhile to investigate what happens in this model when working memory capacity is extremely low.
31 Table 4. Process Analysis – Moderated Mediation
Model 4
Dependent Variable Voice Quality – Originality
Coefficient (β) SE t LLCI ULCI
Constant 0.492 1.919 0.257 -3.297 4.282
Voice Quantity (VQT) 0.712 0.396 1.797 -0.070 1.493
Bright Side Risk-taking 0.105 0.086 1.226 -0.064 0.275
Working Memory Capacity (WMC) 0.171 0.292 0.586 -0.405 0.747
VQT*WMC -0.020 0.062 -0.314 -0.142 0.103
Age -0.008 0.006 -1.353 -0.020 0.004
R2 0.327
Note. N = 166.
Table 5. Process Analysis – Conditional Indirect Effects at Values of the Moderator
Dependent Variable Voice Quality – Originality
WMC Effect SE LLCI ULCI
Voice Quantity (- 1 SD) 5.247 0.115 0.066 -0.004 0.253
Voice Quantity (M) 6.283* 0.111 0.064 0.001 0.255
Voice Quantity (+ 1 SD) 7.320* 0.107 0.068 0.005 0.277
32 Figure 2. Effects Model
Discussion Summary of Study Results
With our study, we intended to add knowledge to the literature about employee voice behavior. More specifically, we studied how risk-taking and cognitive capacity affect the quality of employees’ voice. The research question we tried to answer with our study is as follows: “Do risk-takers learn how to voice original ideas through practice? And does cognitive capacity enhance that process?”
Our exploratory factor analysis revealed some interesting results, especially for employees’ risk-taking propensity. The six dimensions we used based on Nicholson and colleagues (2005: health, financial, safety, recreational, career, and social) fell into two factors that can be described as the bright and dark side of risk-taking. Taking more self-destructive risks, such as taking health, financial, or safety risks, belong to the dark side of risk-taking, whereas taking more adventurous risks, such as taking recreational, career, or social risks, belong to the bright side of risk-taking. This finding is somewhat in line with current literature on risky behavior: taking health risks, such as drug addiction, is seen as self-destructive behavior (Baumeister & Scher, 1988), whereas taking recreational risks, such as rock climbing, is seen as adventurous behavior (Levenson, 1990).
Voice Quantity
Bright Side
Risk-taking Voice Originality
Working Memory Capacity
β = 0.103, 95%CI: -0.064 – 0.271 β = 0.113, 95%CI: 0.000 – 0.258
33 Our first hypothesis stated that employees with high risk-taking propensity are likely to engage in risky behaviors on the work floor, and therefore are likely to communicate ideas, suggestions, problems, and opinions more often compared to employees with low risk-taking propensity. This hypothesis was supported: there is indeed a positive effect of employees’ risk-taking propensity on voice quantity. However, this effect is only present for the bright side risk-taking, and when controlled for age. This means that employees who take more adventurous (recreational, career, and social) risks also communicate ideas, suggestions, problems, and opinions more often. This could be explained by why the decision to voice is seen as a risky decision, especially with regard to career and social aspects of work. Not all supervisors appreciate voice behavior, thus the supervisor-employee relationship might get damaged (social risk), which in turn could result in, for example, a promotion being denied by the manager (career risk).
Our second hypothesis included a mediation effect, stating that employees with high risk-taking propensity are likely to voice more often, thereby practice voice behavior, and therefore are more likely to communicate ideas, suggestions, problems, and opinions that are original and useful compared to employees with low risk-taking propensity. This hypothesis was partially supported: we found an indirect effect of employees’ bright side risk-taking propensity on voice originality, mediated by employee voice quantity. Again, this effect is present when controlled for age. This indicates that employees with high adventurous risk-taking propensity communicate ideas, suggestions, problems, and opinions that are of higher originality because they voice more often compared to employees with low risk-taking propensity. As mentioned before, risk-taking distinguishes radical innovations from incremental ones (Ekvall, 1996), where the latter is more about small but useful improvements and the former is more about big, original, and out of the box changes. This
34 could explain why our results show that employees who tend to take more (adventurous) risks also voice more original output than employees who don’t tend to take risks.
Our third hypothesis included a moderated mediation effect, stating that employees with high risk-taking propensity achieve higher voice quality through voice quantity when they also posses high working memory capacity. Employees with low working memory capacity might not achieve the same quality of voice, even though they also engage in more risky voice behavior. Unexpectedly, we found no support for this hypothesis. The mediation effect was independent of cognitive capacity, indicating that employees who take more adventurous risks communicate higher quality ideas, suggestions, problems, and opinions through voicing more often, regardless of their cognitive abilities.
However, the interaction effect seemed to be present for higher values (M and M + 1 SD) of working memory capacity, indicating that the mediation model is slightly dependent on employees’ cognitive capabilities only when it concerns employees of above average intelligence. Although the indirect effect of employee voice quantity seemed to become significant as the working memory capacity value increases, the strength of the effect decreases. This suggests that employees with high cognitive capacity and high bright side risk-taking propensity depend on this cognitive capacity to reach higher voice originality through practicing voice more often.
Alternative Explanations
An explanation for finding a full mediation effect may be that employees with high risk-taking propensity voice more often, learn to voice more original ideas, suggestions, problems, and opinions, and therefore voice more automatically without the need to overthink things. This line of reasoning aligns with dual process theories, which distinguish between (1) unconscious and automatic information processing and (2) conscious and systematic
35 information processing. According to these theories, the automatic process is independent of working memory capacity, whereas the systematic process is limited by working memory capacity (Evans, 2008; Barrett, Tugade, & Engle, 2004). This explains why we did not find an interaction effect. Doing something over and over again can result in doing something automatically without the need for thinking about it. For example, when you lock your door every morning when you leave for school or work over and over again, it becomes an automatic action you undertake (almost) unconsciously. There is no need to systematically think about it, and thus there is no need for cognitive capacity. Looking at voice behavior, we see that adventurous risk-takers voice more often, seem to practice and enhance their voice, and therefore voice original output. Since these employees voice so often, it might become something they do automatically, without being aware of it. These employees thus seem to take the automatic route in dual process thinking. This differs from our initial explanation of how cognitive capacity might influence employee voice behavior: through systematic thinking and focusing.
An explanation for finding a decreased mediation effect higher values of employees’ working memory capacity might be that employees’ cognitive capacity only influences their voice originality when there is “space” left for complex thinking. Thus, when working memory capacity is so high, that there is always space left to focus on and process relevant information. As stated above, dual process theories argue that people’s systematic information processing is limited by their working memory capacity (Evans. 2008). Therefore, it might be that employees with higher working memory capacity can achieve high voice originality through systematic thinking, whereas employees with lower working memory capacity reach higher voice originality solely through the automatic route, by practicing more. Thus there might be different strategies, or routes, possible for achieving original voice. This “dual route” thought is somewhat in line with the dual pathway to
36 creativity model (De Dreu, Baas, & Nijstad, 2008). According to this model, there are two ways to achieve creativity, or the generation of original ideas, namely cognitive flexibility and cognitive persistence (Nijstad, De Dreu, Rietzschel, & Baas, 2010; De Dreu, Baas, & Nijstad, 2008). As discussed in the theory section, the cognitive persistence route leads to original idea generation through focused attention and systematic thinking, whereas the cognitive flexibility route leads to original idea generations through reduced attention and switching perspectives. Linking this to our moderation model findings, it can be that employees with low working memory capacity generate original ideas, suggestions, problems, and opinions through the cognitive flexibility route, thereby practicing voice so often due to their risk-taking propensity that it becomes an automatic behavior on the work floor. On the other hand, employees with high working memory capacity might generate original ideas suggestions, problems, and opinions through the cognitive persistence route, depending on their cognitive capacity for focused attention and systematic thinking. These employees then might depend on their risk-taking propensity to also voice these original ideas, suggestions, problems, and opinions. Thus, employees can achieve high quality voice by generating original content through focused attention and systematic thinking, whereby they need their cognitive capacity and depend on their risk-taking propensity to also voice these generated ideas. Or employees achieve high quality voice through reduced attention and switching perspectives, whereby they voice more often due to their high risk-taking propensity, and thus voice automatically and do not need their cognitive capacity.
Methodological Points of Critique and Positive Points
Another possible methodological explanation for not fining an interaction effect could be that the participants of this study were sampled through convenience and snowball sampling. Students contacted potential participants who were easily accessible to them. These contacts
37 then asked their colleagues, supervisors, or employees. This could have resulted in an above average intelligent sample. The average number of items participants in our study could keep active in their memory was 6.270 (SD = 1.033), whereas Cowan (2010) argues that working memory capacity in adults is limited to 3-5 items on average. We suggest that this study should be repeated with a sample consisting of randomly chosen participants, so that the working memory capacity levels are more diverse and representative of the general average in the literature.
Furthermore, the same students were responsible for the data collection as well. These students were not trained in data collection, and this could have influenced the data collection process. Fore example, students could have explained the working memory capacity task incorrectly, which could have resulted in participants holding the iPad incorrectly. Moreover, students could have forgotten to ask for a closed office space for this task, resulting in participants being distracted by their colleagues. For future studies we suggest the use of two or three trained researchers for both the sampling and data collection process.
Moreover, the current literature does include risk as a factor that could explain employee voice behavior, this is mostly the perception of risk and not the tendencies towards risk-taking. Our study, however, included a risk-taking tendency measure covering different domains (career, social, recreational, health, financial, and safety) for past and present risk-taking, giving us a better overview of employees’ general tendencies to take risks in their daily lives. Looking to risk-taking tendencies in six different domains allowed us to look into different combinations of risk-taking tendencies.
Suggestions for Future Research
First of all, we suggest looking deeper into the risk-taking behavior of employees. Especially since we found that only adventurous risk-taking propensity seems to influence employees’
38 original voice behavior. It would be interesting to see whether there are some factors that might influence these adventurous risk-taking tendencies. For example, risk-sensitivity theory argues that risk-taking behavior is dependent upon a situation of need (Mishra, 2014; Mishra, Barclay, & Lalumière, 2014; Mishra & Lalumière, 2010). Risk-sensitivity theory reasons that people engage in more risky behaviors when low-risk options cannot satisfy the needs they have in that particular moment (Mishra, 2014). For example, people with low income are more likely to engage in risky behaviors to improve their situation (Wilson & Daly, 1997). Drawing on risk-sensitivity theory, it seems likely that employees with low income might voice more than employees with high income in order to increase their income by increasing their performance. Although we did not find a relationship between dark side (health, financial, and safety) risk-taking propensity, it is interesting to look at how factors such as having a partner or not, or having below or above average income influences employees’ adventurous risk-taking tendencies shown on the work floor, and thus might influence their voice behavior. Especially since risk-taking propensity and risk-sensitivity theory are not each others’ opposite, and might be related to each other in a way that one (risk-sensitivity) explains the other (risk-taking propensity).
Moreover, it would be interesting to look deeper into the different dimensions of risk-taking propensity. More specifically, we suggest looking into the different effects of past and present risk-taking on employee voice behavior since risk-taking behavior decreases with age (Deakin, Aitken, Robbins, & Sahakian, 2004).
Furthermore, as discussed in the alternative explanations section, it is possible that there is not one, but rather two routes employees can take in their journey to achieve original voice. One route is characterized by automatic behavior, whereby employees voice original ideas because they practice voice behavior to the point that it becomes an automatic behavior. The other route is characterized by systematic thinking, whereby employees voice original
39 ideas because they have the ability to focus on and process relevant information. This thinking aligns with dual process theories, where there is a distinction between automatic versus systematic thought processes (Evans, 2008). One dual process theory is specific to creativity, or the process of generating original ideas (Amabile, 1988), namely the dual pathway to creativity model (Nijstad, De Dreu, Rietzschel, & Baas, 2010; De Dreu, Baas, & Nijstad, 2008). It would be interesting to look at these two different routes in relation to risk-taking tendency. It might be that employees who tend to take risks take the flexibility route to achieve creativity, especially those with adventurous risk-taking tendencies, since they might adventurously switch from one perspective to another more easily.
Moreover, since the communication of original ideas might be the bridge between idea generation (creativity) and idea implementation (innovation), we suggest looking to the role of creativity as a moderator on the relationship between risk-taking and voice quality (originality).
Interpretations and Contributions
In alignment with our expectations, our results show that employees’ voice quantity indeed has a mediating effect in the relationship between employees’ bright side risk-taking propensity and voice originality. However, we did not find that this mediation model is dependent on employees’ working memory capacity. In other words, employees who have a high tendency towards adventurous risk-taking, voice more often, and therefore communicate more original ideas, suggestions, problems, and opinions compared to those with a low tendency towards adventurous risk-taking. Moreover, they do so without the need for using their cognitive abilities for complex thinking and information processing.
Our study’s contribution to the literature is threefold. First, by making a distinction between the quantity and quality (originality) of voice in both theory and measurements, we
