Risk-taking behavior, attention control and
voice quality
Does an employee’s (situational) ability to control attention moderate the
relationship between an employee’s risk-taking tendencies and the
quality of ideas the employee communicates towards supervisors?
8.5
Name: Gizem Özmen
Student number: 10010823 Supervisor: Inge Wolsink
Faculty: Faculty of Economics and Business University: University of Amsterdam
Date: July 11, 2014
Abstract
Previous voice behavior studies have focused on the frequency of voice, rather than the quality (Wolsink, forthcoming). This study focuses on the quality of voice, and the effect of attention control and risk-taking tendency. The reason for choosing voice quality is that organizations might improve more when employees focus on problems and solutions, and communicate innovative ideas, than when employees just suggest the first simple things that arises into their mind. We assume that creativity is related to the quality of voice, and since attention control seems to benefit creativity (De Dreu, Baas, Nijstad, Wolsink and Roskes, 2012), we hypothesize that the ability to control attention positively relates to voice quality. Furthermore, previous studies showed that a perception of high risk negatively relates to voice behavior (Detert & Burris, 2007; Lui, Zhu and Yang, 2010; Milliken et al., 2003). Therefore, we hypothesize that risk-taking tendency positively relates to voice quality. Forty-eight participants took part in our experiment, where we manipulated their attention control with low and high working memory loads (within participants), measured their voice quality (within participants, evaluated by two independent judges), and their risk-taking tendency (between participants, BART task). We conclude that attention control indeed has a positive effect on voice quality, while the effect of risk-taking tendency on voice remains unknown.
Introduction
Nowadays, organizations rely more and more on the input of their employees. For example, some frontline employees and a store manager of a Starbucks store in Southern California came up with the idea of the Starbucks Frappuccino, now a billion-dollar product (Smith, 2012). Furthermore, Starbucks also launched a forum, called My Starbucks Idea, where employees (and customers) can share their ideas to improve Starbucks and its products.
This phenomenon of proactive voice behavior, whereby employees communicate ideas to their supervisors (Van Dyne & LePine, 1998; Detert & Burris, 2007), has been researched extensively. However, extant literature focuses only on how often (frequency of voice) employees communicate their ideas, and do not measure the usefulness and originality (voice quality) of these communicated ideas (Wolsink, forthcoming). Therefore, the focus of this study will be on the quality of employee voice, thus the communication of useful and original ideas.
Little research has been conducted that addresses whether voice behavior actually contributes to organizational success (Morrison, 2014). However, there are two empirical studies that show a positive effect of voice behavior on the (1) financial performance of the organization (MacKenzie, Podsakoff & Podsakoff, 2011), and (2) organizational unit effectiveness (Detert, Burris, Harrison & Martin, 2013).
Voice behavior of employees can be influenced by a variety of factors. For example, rational thinking suggests that the cognitive idea generation abilities of an employee can determine the usefulness and originality of the generated ideas. Previous research has shown that an employee’s creative thinking can be affected by his/her ability to control attention (working memory capacity) and focus fully on the idea generation process (De Dreu, Baas, Nijstad, Wolsink & Roskes, 2012). Which is why we intent to investigate whether attention facilitates the quality of voice as well.
Furthermore, risk-taking behavior of employees might also influence the communication of ideas. Employees are more reluctant to communicate their ideas towards their supervisors when they perceive the risk of speaking up to be high (Milliken, Morrison & Hewlin, 2003). So, when employees have low levels of perceived risk or high risk-taking tendencies in general, they might make different risk-calculations for communicating ideas towards supervisors. However, the opposite may be true when ideas of good quality are concerned. Rational thinking suggests that employees are more likely to communicate ideas towards supervisors when they are high risk-takers, and because they communicate more frequently, chances are higher that they also communicate high quality ideas and that, by practicing more than non-risk takers and getting more feedback, they increase their voice quality over time.
As stated above, there is a difference between the frequency and quality of voice, and it is likely that the quality of voice has more value for organizations than the frequency of voice. Organizations might improve more when employees focus on problems and solutions, and communicate innovative ideas, than when employees just suggest the first simple things that arises into their mind. Thus, it is important for organizations to know the difference between the frequency and quality of communicated ideas, and to know how the risk-taking behavior and attention control abilities of employees can affect both components of voice behavior.
The purpose of this study is to see whether the communication of useful and original ideas (voice quality) is dependent of an employee’s ability to control attention, and to find out whether the risk-taking behavior of the employee increases
the strength of this relationship. In other words, is the quality of ideas communicated to a supervisor highest when employees have high attention control and have a tendency to take risks? To fulfill this purpose, the following research question will be answered:
Does an employee’s (situational) ability to control attention moderate the relationship between an employee’s risk-taking tendencies and the quality of ideas the employee communicates towards supervisors?
Theoretical Framework Voice Behavior
Voice behavior is the communication of innovative ideas, suggestions, concerns about problems, opinions, and improvement possibilities within the organization, even when others disagree (Van Dyne & LePine, 1998; Detert & Burris, 2007; Morrison, 2011; Tangirala & Ramanujam, 2008). Although the quality of voice is mentioned in the definitions of voice behavior (see the definition above: communication of innovative ideas), most studies only measure the frequency of voice. An example of such measurement is the study of Van Dyne and LePine (1998). They used a questionnaire including the following item: “This particular co-worker speaks up in this group with ideas for new projects or changes in procedures”. This is a measure of voice frequency rather than quality.
Furthermore, voice behavior is seen as an extra-role behavior, which means that communication of innovative ideas and suggestions is not part of the primary job requirements (in-role job tasks) (Van Dyne & LePine, 1998). Employees who voice do it voluntarily.
The current literature splits voice behavior up in two kinds: speaking out and speaking up, whereby the latter is to communicate ideas to supervisors and the first is to communicate ideas to colleagues (Lui, Zhu and Yang, 2010; Detert & Burris, 2007; Morrison, 2014). In this study, we focus on speaking up; communicating ideas towards supervisors. Another distinction that scholars make is that voice behavior can be suggestion-focused, problem-focused or opinion-focused. Opinion-focused voice concerns opinions that are different from others within the organization, problem-focused voice contains concerns about the organization, and suggestion-problem-focused voice includes ideas and suggestions for improvements within the organization (Morrison,
2011). All three focuses can be found in the above definition of voice. However, in this study we will only focus on the communication of (useful and original) ideas for improvement within the organization, thus focus on suggestion-focused voice. It seems logical that creativity, and thus attention control, is more strongly related with suggestion-focused voice, and less with opinion- or problem-focused voice. The latter two are less about finding solutions and creating new situations, and therefore less in need of attention.
Current literature assumes that employees make conscious deliberations on whether or not to communicate their ideas towards supervisors (Morrison, 2011). Presumably, one should need attention and attention control to do this. However, whether employees indeed consciously calculate risks and rewards, has not yet been empirically tested to this date. Therefore, we will examine the influence of attention control on employee voice behavior, more specifically on the quality of voice.
Voice Quality, Creativity and Attention Control
Existing literature about voice behavior only addresses how often employees communicate ideas towards their supervisors. Research investigating the usefulness and originality of these communicated ideas is scarce (Wolsink, forthcoming).
Although most literature only measures the frequency of communicated ideas, there is one study (Grant, 2013) that tried to address the quality of voice by measuring the constructiveness (usefulness) of ideas and suggestions. However, usefulness alone is not enough to measure the quality of ideas. For organizations to benefit from employee voice, ideas also need to be innovative, and therefore original and creative. Therefore, we define voice quality as the usefulness and originality of communicated ideas.
Usefulness and originality of ideas is generally referred to as creativity (Wolsink, forthcoming). The creative abilities of an employee depend (partially) on his or her ability to control attention (De Dreu, Baas, Nijstad, Wolsink and Roskes, 2012). Attention control and persistence result in being focused on a goal, the generation of a creative idea, and will lead to more original ideas. It is likely that being able to concentrate attention will result in combining more information and less distraction by non-relevant things, and therefore employees will voice less random ideas and more useful ideas that will benefit the organization. Thus the ability to concentrate on the goal, knowing what is important to communicate, and not being
afraid to communicate this could result in a higher voice quality, since employees will generate, and therefore voice more original ideas.
The Positive Effect of Risk-taking Behavior on Voice Quality
Employees can be reluctant to communicate ideas towards their supervisor when they perceive the risk of challenging the status quo within the organization to be high (Detert & Burris, 2007; Lui, Zhu and Yang, 2010; Milliken et al., 2003). Communicating ideas can result in a negative image (Ashforth & Humphrey, 1995) or in a damaged relationship between employees and their colleagues and supervisors (Adler & Kwon, 2002). Therefore, a perceived high risk will stop an employee from communicating ideas, which is referred to silence in the literature (Morrison, 2014), even if an employee’s cognitive abilities allow him or her to generate useful and original ideas. However, a perceived high risk will not stop an employee from speaking up if the employee in question has high risk-taking tendencies. Therefore, we assume that employees with high risk-taking tendencies will voice more often than employees who take none to low risks.
Current literature defines voice behavior as the communication of innovative ideas towards supervisors. However, only the frequency of voice is being measured, excluding the quality of voice (Wolsink, forthcoming). This distinction between the frequency and quality of voice behavior is a new concept, and therefore there is no proof of a relationship between the frequency and quality of voice behavior yet. However, rational thinking suggests that an idea, no matter how great the idea, can only be useful when an employee communicates this idea.
As stated above, current research suggests that the frequency of voice can be influenced by a variety of factors, for example, leadership style (Detert & Burris, 2007; Lui, Zhu & Yang, 2010) and the risk of speaking up (Detert & Burris, 2007; Lui, Zhu and Yang, 2010; Milliken et al., 2003). Thus, employees who have high risk taking tendencies are more likely to voice, and employees who voice more often are more likely to communicate useful and original ideas. Therefore, we hypothesize that:
H1: Employees with high risk-taking tendencies are more likely to voice, and therefore are more likely to communicate the useful and original ideas they have generated.
The Moderating Effect of Attention Control
Attention control, also known as executive control, is an individual’s ability to control attention while doing different tasks (Kane & Engle, 2002; Kane & Engle, 2003). The ability to focus attention on information that is relevant is a requirement for creative thinking (Dietrich, 2004). Central to attention control is working memory capacity (Kane & Engle, 2002; Kane & Engle, 2003).
Working memory capacity is defined as the central stage of higher-order cognition, which is associated with cognitive functions such as effective reasoning (Süß, Oberauer, Wittmann, Wilhelm & Schulze, 2002), decision-making and action planning (Oberauer, Süß, Wilhelm & Wittman, 2003). Working memory capacity helps people with processing available information (Oberauer & Bialkova, 2009) in a conscious and controlled way (Evans, 2008) and is related to hypothetical thinking about future opportunities, and creative insight (De Dreu, Baas, Nijstad, Wolsink and Roskes, 2012). Putting a load on working memory (working memory load), by distracting attention, can influence one’s ability to control attention. The higher the load on working memory, the lower one’s attention control. In their study, De Dreu, Baas, Nijstad, Wolsink and Roskes (2012) found that participants with a low load on their working memory performed better on a creative task than participants with a high load on their working memory. Thus, attention control and working memory load are related to the generation of creative ideas. However, little research addresses the effect of attention control and working memory load on the communication of creative, and therefore original and useful ideas.
Scholars argue that there are two ways to achieve creativity: cognitive flexibility and cognitive persistence (De Dreu, Baas & Nijstad, 2008). Cognitive flexibility is associated with reduced attention control, whereby individuals can easily switch to different perspectives and approaches. Cognitive persistence is related to systematic thinking and attention focus (Nijstad, De Dreu, Rietzschel & Baas, 2010).
Employees with high attention control will have high cognitive persistence and will be able to focus their attention on relevant information. Therefore, they will be better able to select and communicate only the highest quality of ideas. We hypothesize that:
H2: Employees with high risk-taking tendencies will voice more often, and communicate more useful and original ideas when they have high attention control.
Figure 1. Conceptual Framework
Method Design and Participants
Fulltime or part-time students (N = 71, 46.4% female) of the Faculty of Economics and Business (University of Amsterdam) participated for €7.50 - €15, and were randomly assigned to either scenario 1 (first high than low working memory load) or scenario 2 (first low than high working memory load). The dependent variable was the number of useful and original ideas communicated. The independent variables were attention control, a within-subject variable, and risk-taking behavior, a between-subject variable. We hoped to achieve a good representation of the sexes within the Dutch work force, which consists of approximately 46.5% female and 53.5% male employees (Centraal Bureau voor de Statistiek, 2013, June 25).
We removed three participants who had a SD of 0.00 for their risk-taking measure, because it is likely that they did not take the task serious. Furthermore, we filtered out 20 participants who did not completed one or both of the working memory load tasks in the correct way. We did this for the reason that we could not control if our manipulation succeeded. This resulted in a final N of 48 participants (average age = 22.11, SD = 1.45, 40.4% female). Unfortunately, this did not result in a good representation of the Dutch work force.
Measurements, Tasks and Manipulations
Dependent Variable Voice Quality and the Voicing Task
We measured voice quality with a voicing task developed by Wolsink (forthcoming). First, participants had to generate 10 ideas about improvements for the Faculty of Economics and Business and the university in general. This idea generation task, or creativity task, is adapted from De Dreu, Baas and Nijstad (2008). We used this
Attention Control
Voice Quality Risk-taking Behavior
Hypothesis 1
creativity task, because creativity is linked to quality of voice. Second, participants had the opportunity to voluntarily communicate ideas towards the experiment leader, and choose the ideas they would like to communicate. We chose for voluntary communication of ideas because voice behavior is seen as an extra-role behavior. This means that communicating ideas is not a requirement of the job. Two raters rated the communicated ideas on a 1-5 Likert scale, where a score of 1 stands for not useful and original and a score of 5 stands for very useful and original. In other words, participants who score high (4 or 5) have communicated ideas that are useful and original. The measurement of voice quality was the score of the best idea communicated.
Independent Variable Risk-taking Behavior and the Balloon Analogue Risk Task We measured risk-taking behavior with the Balloon Analogue Risk Task (BART) (Lejuez, Read, Kahler, Richards, Ramsey, Stuart & Brown, 2002). In this task, participants are asked to pump up a balloon in return for money. They keep their earned money if the balloon does not explode. Participants are told that on average, the balloon explodes after 64 pumps. Risk-taking tendency is the direct sum-score of the amount of times the participant pumps up the balloon. Participants repeated this procedure 40 times. The measurement of risk-taking behavior was the average amount of pumps.
Manipulation: Moderating Variable Attention Control and the Working Memory Load Task
We manipulated the attention control of participants with both a high and a low working memory load condition, whereby the attention of the participant was distracted. We did this to see whether distracting attention influences the communication of high quality ideas. In the high load condition participants needed to count backwards from 107 in steps of 3 (107, 104, …), taking each step after three seconds when they hear a “beep” sound. In the low load condition participants needed to count backwards from 100 in steps of 1 (100, 99, …), again, taking each step after three seconds when they hear a “beep” sound. To check our manipulation, participants practiced the working memory load task before generating 10 ideas, where they counted back from 100 in steps of two. Furthermore, participants needed to count vocally, so that the experiment leader could hear the participant counting
while selecting ideas, and participants also needed to fill in the correct number they ended on. Participants either had high attention control (low working memory load) or low attention control (high working memory load).
Control Variables
We controlled for commitment to the University of Amsterdam, since this might influence participants’ performance on the voicing task, more specifically, the frequency or quality of ideas participants generate and communicate to improve the university. Participants who are more involved with the university are more likely to take this task serious, and also probably already had some ideas for improvement on their mind. Commitment was measured with a set of five questions (see Appendix A). An example question is: ‘To what extent do you feel involved in the Faculty of Economics and Business?’, on which participants can reply with a 1-5 Likert scale ranging from (1) in a very slight degree to (5) in a very strong degree. The higher the score, the more involved/committed a participant is.
Figure 2. Measurement Model
Balloon Analogue Risk Task (Lejuez, Read, Kahler, Richards, Ramsey, Stuart &
Brown, 2002) Average amount of pumps
Voicing Task (Wolsink, forthcoming) Highest score for originality and
usefulness of best idea communicated
Working Memory Load Task
(Wolsink, forthcoming) High vs. low load
Attention Control
Voice Quality Risk-taking Behavior
Hypothesis 1
Procedure
A week before their appointment, participants received a pre-test, which contained some demographical and commitment questions. On the day of the appointment participants came to the laboratory and were seated in an office space equipped with a chair, a desk, a computer with two screens, two keyboards and two mice. They were told that they would participate in a study on attention and multi-tasking. Instructions were given on the computer. Participants were randomly assigned to either scenario 1 (first the high than the low load condition) or scenario 2 (first the low than the high load condition).
Participants first were told that they had to, anonymously, generate 10 ideas for improvement on the Faculty of Economics and Business. We chose this anonymous brainstorm, or creativity task, (De Dreu, Baas and Nijstad, 2008), because high quality ideas are creative ideas. The experiment leader rated the ideas immediately after participants generated the ideas. Ideas received either code 1, which stands for not useful and original idea, or code 2, which means useful and original idea. The computer than assigned the ideas to either the high or low load condition, each condition containing five ideas and preferably the same amount of code 1 (bad) ideas. This was done in order to make sure that high quality ideas were equally represented in both high and low working memory load conditions.
Following the idea generation task, participants were explained that they would get to see their own ideas, and that they had 30 seconds to pick any ideas they would like to, voluntarily, communicate to the board of their faculty. However, any idea they would communicate, would be judged by the experiment leader, who would decide whether the idea was good enough to pass on to the board of the faculty. Risk of idea communication was increase by giving a reward (+ €0.50) or decrease by punishment (- €0.50) for each communicated idea, based on the quality as rated by the experiment leader. During the decision phase of voice, participants had to complete their primary, or in-role, task, which was the load task and represented the required job tasks of employees. To make sure their primary goal would be the load task, we rewarded them with €2.50 upon correct completion. Voicing was therefore their secondary, or extra-role, task, which represented non-required voluntary job tasks of employees. The voicing task was executed twice, once with a low working memory load and once with a high working memory load, in randomized order. Hence, the quality of ideas communicated could be determined within participants, ruling out any effects of
individual differences in creativity.
The last task participants completed was the BART task, where participants had to inflate 40 balloons by filling in the number of pumps by each balloon. Participants could earn €0.005 per pump, when the balloon in question does not blow up.
After finishing the BART task, participants were asked to fill in an exit questionnaire consisting of four questions about the experiment leader (see Appendix B) and received the amount that they had earned with the tasks.
Analyses and Predictions
We used a mixed design repeated-measures ANOVA to test the main effects and the interaction effect on our dependent variable voice quality, since the study had a mixed design of both a within- (working memory load) and a between-subject (risk-taking) independent variable.
Main Effects
We predicted a positive main effect of risk-taking behavior on voice quality, which means that we expected to find that participants with high risk-taking tendencies would also communicate more useful and original ideas.
Furthermore, we predicted a positive main effect of attention control on voice quality. In other words, we expected to find that participants with high attention control would also be able to voice the highest quality ideas.
Figure 3. Main Effects
Interaction Effect
We predicted that attention control positively moderates the relationship between risk-taking behavior and voice quality. In other words, we expected that participants who have high risk-taking tendencies would voice more often in general, and if their
Attention Control
Voice Quality Risk-taking Behavior
attention control was high (thus in the low working memory load condition), they would communicate ideas of higher quality than participants with low attention control (thus in the high working memory load condition).
Figure 4. Interaction Effect
Results Reliabilities and Correlations
We conducted a reliability analysis for the dependent variable voice quality. Two raters scored the ideas for voice quality on two aspects: originality and usefulness. We conducted an inter-rater reliability test to see whether the raters gave more or less the same scores to ideas. The inter-rater reliabilities were reasonable: 0.67 for originality and 0.54 for usefulness, thus we took the average of both rater scores as our measure of voice quality.
Furthermore, we conducted a correlation analysis (see Table 1). Unexpectedly, the analysis showed no correlation between the three main variables. An additional correlation analysis (see Table 1) including the voice frequency of participants showed that there is a stronger correlation between risk-taking behavior and the voice frequency in the high attention control (low working memory load) condition than in the low attention control (high working memory load) condition. Both correlations are not significant, however, the correlations show a trend where risk-taking tendencies have a greater effect on the voice frequency of participants with high attention control than participants with low attention control. In other words, participants with high attention control tend to voice more often when they have high risk-taking tendencies, while participants with low attention control will show less increase in voice frequency when they have high risk-taking tendencies. It can be that the power of our data is not high enough to show this effect, however, it offers opportunities for future studies on this topic. Additionally, the results showed a positive correlation between
Attention Control
Voice Quality Risk-taking Behavior
voice frequency and voice quality in the high attention control condition. However, there is no correlation between frequency and quality in the low attention control condition. This indicates that the relationship between frequency and quality is much stronger (r = 0.5, p < 0.001) when people are mildly distracted than when people are heavily distracted (r = -0.07, ns). This means that when people are able to control their attention, people who communicate many ideas also communicate the best ideas, and people who communicate few ideas voice ideas that are not original and useful. Presumably, attention control provides people with the tool to be aware of their own creative capabilities, hence, it helps them evaluate the quality of their ideas, which than results in either voicing or not. However, when attention drops, thus when the working memory load is high, either people with low quality ideas loose the ability to monitor their own mistakes and voice more frequently than they should, looking at their creative capabilities, or the people with high quality ideas loose the ability to select the best ideas and therefore voice less than they should, based on their creative capabilities.
This additional correlation analysis, which also included the student type of participants, showed a negative correlation between student type and voice quality in the low attention control condition. However, we cannot draw any conclusions based on this correlation, since this is a nominal variable and the participants are not evenly distributed within the two categories (N fulltime students = 33, N part-time students = 5).
Table 1. Descriptives and correlations between variables
M SD 1 2 3 4 5 6
1 Voice Quality (DV) in Low Attention Control (MV)
condition a 2.202 0.400 1 0.267 -0.039 -0.070 -0.157 -0.320* 2 Voice Quality (DV) in High Attention Control (MV)
condition b 2.614 0.606 0.267 1 -0.117 -0.009 0.500** -0.091 3 Risk-taking Behavior (IV) d 60.576 9.710 -0.039 -0.117 1 0.181 0.200 -0.099 4 Voice Frequency in Low Attention Control condition d 2.250 1.523 -0.070 -0.009 0.181 1 0.215 0.014 5 Voice Frequency in High Attention Control condition d 2.520 1.544 -0.157 0.500** 0.200 0.215 1 -0.180 6 Student type c 1.13 0.337 -0.320* -0.091 -0.099 0.014 -0.180 1
Repeated-measures ANOVA Main Effects
As predicted, the repeated-measures ANOVA showed that attention control indeed has a positive main effect (F (39) = 18.092, p < 0.001, η = 0.328)on the quality of voice, with an observed power of 0.985 (see Table 3). This indicates that participants with high attention control, thus in the low working memory load condition, communicated ideas of higher quality than participants with low attention control, thus in the high working memory load condition. In other words, an inability to control one’s attention seems to affect the ability to select the highest quality ideas. Therefore we conclude that attention control is important for voice quality.
Unexpectedly, the repeated-measures ANOVA did not confirm our prediction that risk-taking behavior would have a positive main effect (F (39,1) = 0.000, ns, η = 0.000) on the overall quality of voice, with an observed power of 0.050 (see Table 3). This means that participants with high risk-taking tendencies did not communicated ideas of higher quality than participants with lower risk-taking tendencies.
Interaction Effect
Unfortunately, the repeated-measures ANOVA did not confirm our prediction that there is an interaction effect (F (39) = 0.464, ns, η = 0.012) of risk-taking behavior and attention control on voice quality, with an observed power of 0.102 (see Table 3). This indicates that participants with high risk-taking tendencies do not have a higher quality of voice when their attention control is high in comparison to participants with low attention control.
However, the descriptives of the repeated-measures ANOVA showed some interesting results (see Table 2). The results implicate that there might be an interaction effect. In the low attention control condition, participants had a higher mean value for voice quality in the high risk-taking behavior group (M = 2.121, SD = 0.379) than in the low risk-taking group (M = 2.190, SD = 0.376). The results showed the opposite for the high attention control condition, where the mean value for voice quality is lower in the high risk-taking behavior group (M = 2.550, SD = 0.652) than in the low risk-taking behavior group (M = 2.618, SD = 0.580). However, this effect, if it does exist at all, is very small and not detectable with the power of our data.
Figure 6. Conceptual model of the variables and their Partial Etas
Additional repeated-measures ANOVA
We conducted an additional repeated-measures ANOVA to further investigate the effect of student type on our model, because it is likely that students who are at the faculty longer or more often are better trained in thinking about ideas for improvement or already have some ideas for a longer period, and thus suffer less from the working memory load manipulation. The results showed that within the part time student group the main effect of attention control on voice quality was stronger (η = 0.760) than within the group of fulltime students (η = 0.318) (see Table 3). However, these findings do not hold any statistical significance, since the distribution of part-time (N = 5) and fullpart-time (N = 33) students is not even.
2 2,2 2,4 2,6 2,8
Low Attention Control High Attention Control
V o ic e Q u ality High Risk-taking Behavior Low Risk-taking Behavior Attention Control Voice Quality Risk-taking Behavior (η = 0.000, ns) (η = 0.328, p < 0.001) (η = 0.012,ns)
Table 2. Repeated-measures ANOVA Descriptives
Risk-taking Behavior
Voice Quality
M SD N
Low Attention Control
Low 2.121 0.379 19 High 2.190 0.376 20 Total 2.156 0.374 39
High Attention Control
Low 2.618 0.580 19 High 2.550 0.652 20 Total 2.583 0.611 39
Note. Valid N = 39.
Table 3. Repeated-measures ANOVA Results
Dependent Variable: Voice Quality F η p Power Attention Control 18.092 0.328 0.000 0.985 Risk-taking Behavior 0.000 0.000 0.998 0.050 Attention Control * Risk-taking Behavior 0.464 0.012 0.500 0.102 Attention Control in Part-time Student Group 9.510 0.760 0.054 0.553 Attention Control in Fulltime Student Group 14.454 0.318 0.001 0.957
Note. Valid N = 39.
Summary of Study Results
We intended to add insight into the predictors of quality to the voice behavior literature. More specifically, we studied the effect of risk-taking behavior and attention control on the quality of voice behavior. With our study, we tried to answer the following research question: “Does an employee’s (situational) ability to control attention moderate the relationship between an employee’s risk-taking tendencies and the quality of ideas the employee communicates towards supervisors?”
Our first hypothesis, which states that employees with high risk-taking tendencies are more likely to voice, and therefore are more likely to communicate the useful and original ideas they have generated, is not confirmed. We find no proof for an effect, positive or negative, of risk-taking behavior on voice quality. This means that employees with high risk-taking tendencies will not have a higher voice quality than employees with low risk-taking tendencies. However, as mentioned above in the results section, we find a not significant correlation between risk-taking behavior and voice frequency that is stronger when participants have high attention control.
Our second hypothesis, which states that employees with high risk-taking tendencies voice more often, and communicate more useful and original ideas when they have high attention control, also is not confirmed. The results show no proof of an interaction effect of risk-taking behavior and attention control on voice quality. This indicates that employees with high risk-taking tendencies will not have a higher voice quality when they also have high attention control, than when they have low attention control. However, the results show a significant positive main effect of attention control on the quality of voice. In other words, when employees are under low working memory load, thus are not or mildly distracted by their primary task, they communicate higher quality ideas than when they are under high working memory load, thus when they are heavily distracted.
Alternative explanations
An explanation for not finding a positive effect of risk-taking behavior on voice quality can be that risk-taking is more associated with whether or not to voice and with the frequency of voice rather than with the quality. However, additional analysis shows no significant effect of risk-taking behavior on voice frequency. Not finding any significant effects of risk-taking behavior on both the quality and frequency of voice can be a result of the low power of our study. It is possible that a future study
with a greater amount of participants, and therefore a higher power, will show significant effects.
As mentioned earlier, our results show that risk-taking behavior tends to be more strongly correlated with voice frequency when participants are in the low working memory load condition (high attention control). However, we did not find an interaction effect of risk-taking behavior and attention control. This could suggest that employees with high attention control could think about the risks of speaking up, and thus take risks and voice more systematic. However, it is important to mention that these findings can be a coincidence and further research into the effect of working memory load is needed. Our suggestion is to look at the difference between low working memory load and no load at all. It is possible that a low working memory load condition has the same effect as a no load condition, and that a high load condition does not leave any space for thinking about risks and processing relevant information.
Furthermore, our results show that attention control has a stronger effect on voice quality within the part-time student group than in the fulltime student group. It is likely that most fulltime students probably already had some ideas for improvement of the faculty. Therefore, they could have selected ideas more automatic, for which they did not need attention control. This suggests that there is a learning aspect of voice behavior, the longer or more often employees work at the company, the more they will think about improvements and the more automatic they will voice ideas of good quality.
Methodological Points of Critique
A methodological explanation for not finding the hypothesized effects can be the relatively small sample of 71 participants, which we narrowed down to 48 participants due to missing or incorrect data. The small size of the sample mostly affected our risk-taking and voice quality data. The final sample contained participants who did not voice in one of the conditions, which resulted in an even lower sample for voice quality data (valid N = 39). Furthermore, our risk-taking behavior variable was a between-subject variable, thus the small sample also resulted in a low power for this variable.
Another point of critique regarding our sample is that our sample consisted of students instead of employees, which is not a good representation of voice behavior
within an organization. Furthermore, most of the participants knew one of the experiment leaders, which could have biased their performance on the tasks. It can be that participants did not take tasks seriously and only took part in the study to please the experiment leader they knew. Also, since most participants knew one of the experiment leaders, the experiment leaders were not perceived as an authority figure, and therefore, it is possible that participants voice more easily.
Furthermore, the risk of voice only consists of financial risk. Participants have either a €0.50 bonus or loss. Financial risk is not the only type of risk present in the working environment, risk of a negative image (Ashforth & Humphrey, 1995) or a damaged relationship (Adler & Kwon, 2002) are also perceived risks of voice behavior. However, it is likely that financial risk outweighs the other types of (social) risk.
Positive Points
The topic of this study is new to the existing voice literature. We made a distinction between the quality and the frequency of voice. We focused on the quality of voice rather than the frequency, which existing literature already covered. This distinction resulted in a new finding within the voice behavior topic, namely that an employee’s attention control has a positive effect on his or her voice quality.
Furthermore, our experimental research design is argued to be better than a survey design, since the latter results in a simple correlation and the first results in causality. Also, surveys tend to be more biased than experimental manipulations.
Last, in our study, we measure the quality of voice by looking at the best idea communicated. We chose this measure, because within organizations it is important to get a top idea to the boss. So looking at the score for best idea communicated gives a better representation of voice behavior within organizations. Looking at the average quality of voice, instead of the best idea communicated, could result in employees who only communicate ideas of average quality and none of high quality. It can be that employees who voice a high quality idea also voice a bad idea once in a while, and if we look at the average score of voice quality, these employees’ will be reviewed the same way as employees who only voice average ideas.
As mentioned earlier, the sample of students instead of employees, whom also are acquaintances of the experiment leader, could lead to biased, non-generalizable results. Knowing the experiment leader could result in participants not taking tasks seriously. Furthermore, since the experiment leaders probably were not really perceived as authority figures, it also decreased the risk of speaking up. It is likely that voicing ideas was easier for these participants who knew the experiment leader. For a future experiment we recommend a sample consisting of employees instead of students, and experiment leaders who show authority and are not acquaintances of the participants.
Furthermore, we think that the perceived risk of voice was not clear, or even equal to the risk within real organizations. Our experiment only contained financial risk, however, in organizations voice behavior has also social risks, like a damaged image or relationship. Therefore, in a future experiment, adding social risk to the voicing task could increase the perceived risk of voicing to better represent voice behavior within organizations.
Interpretations and Contributions
In contrast to our predictions, our results show that, in our sample, risk-taking behavior has no effect on voice quality. This means that risk-taking tendencies of employees have no effect on the quality of ideas they communicate. Furthermore, we find no correlation between risk-taking behavior and voice frequency. This indicates that employees with high risk-taking tendencies will not voice more often than employees with low risk-taking tendencies. Our findings contradict the current literature, which argues that risk has an effect on the voice behavior of employees. This could indicate that our measurement of risk is different from that of current literature (which we already discussed above). However, these findings are based on our relatively small sample and low power, and therefore we cannot be certain that there is indeed no effect of risk-taking behavior. Increasing the sample size, and thereby the power, could result in different findings, which could be in line with the current literature.
As predicted, we did find a positive effect of attention control on voice quality. In other words, employees with high attention control will communicate better ideas than employees with low attention control. This is in line with the literature, which states that attention control leads to creativity.
Our study contributes to the literature in twofold. First, we contribute to the voice behavior literature by making a distinction between the frequency and quality of voice. Specifically, where the existing literature focuses on the frequency of voice, we focus on the effect of attention control on the quality of voice. Second, our study contributes to the dual process theory literature, making a distinction between automatic and systematic voice behavior.
Our managerial contribution is also twofold. Our first contribution is specific to the workload of employees. Managers who want their employees to communicate top ideas or suggestions for innovative improvements need to keep the workload of their primary task on a low level. This will ensure a low working memory load, and thus result in employees with high attention control. This way, employees can focus on the relevant information regarding a problem and generate, and communicate, good innovative ideas and suggestions for improvements. Our second contribution is specific to employee recruitment. Managers who are looking for starters who are likely to have high voice quality can benefit from working memory capacity tests. Starters with high working memory capacity test scores are more likely to have high attention control, since working memory capacity is central to attention control, and therefore are more likely to voice ideas of high quality. However, if managers are looking for employees with work experience, they benefit more from recruiting older people with more experience in organizations.
Conclusion
Existing literature on voice behavior argues that voice behavior is the communication of innovative ideas. However, the only measure of voice behavior is the frequency of voice and does not include the quality. In our study we made this distinction clear and focused on the quality of voice.
Furthermore, the existing literature also argues that risk-taking is positively related to voice behavior. However, we found no correlation between risk-taking and either the frequency of quality of voice.
We did find a positive effect of attention control on voice quality. This means that employees with high attention control can focus on the relevant information, generate innovate ideas, and communicate these high quality ideas. Furthermore, we found that attention control tends to have a greater effect on voice quality when employees are new to the organization, than when they work longer or more often in
the organization. However, this cannot be proven with our sample and needs to be researched further. Looking back at the example of voice behavior above, it is possible that the employee who came up with the idea of the Starbucks Frappuccino spent a lot of days thinking of a new drink and focusing on relevant information (such as which ingredients go good together and which not) and finally found the right combination of ingredients, and afterwards communicated this idea.
References
Adler, P. S., & Kwon, S. (2002). Social capital: Prospects for a new concept.
Academy of Management Review, 27, 17−40.
Ashford, S. J., Rothbard, N. P., Piderit, S. K., & Dutton, J. E. (1998). Out on a limb: The role of context and impression management in selling gender-equity
issues.Administration Science Quarterly, 43, 23−57.
Ashforth, B. E., & Humphrey, R. H. (1995). Labeling processes in the organization: Constructing the individual. In L. L. Cummings & B.W. Staw (Eds.),
Research in Organizational Behavior (pp. 413−461). Greenwich, CT: JAI
Press.
Centraal Bureau voor de Statistiek. (2013, June 25). CBS Statline - Arbeidsrekeningen; arbeidsvolume naar bedrijfstak en geslacht. Retrieved
April 30, 2014, from CBS Statline: http://statline.cbs.nl/StatWeb/publication/?DM=SLNL&PA=81108NED&D1=
a&D2=1-2&D3=1&D4=0&D5=1,11,21,31,36-43&HDR=T,G1,G2&STB=G3,G4&VW=T
Crant, J. M. (2000). Proactive behavior in organizations. Journal of management, 26(3), 435-462.
De Dreu, C. K., Baas, M., & Nijstad, B. A. (2008). Hedonic tone and activation level in the mood-creativity link: toward a dual pathway to creativity model.
Journal of personality and social psychology, 94(5), 739.
De Dreu, C. K., Nijstad, B. A., Baas, M., Wolsink, I., & Roskes, M. (2012). Working memory benefits creative insight, musical improvisation, and original ideation through maintained task-focused attention. Personality and Social Psychology Bulletin, 38(5), 656-669.
Detert, J. R., & Burris, E. R. (2007). Leadership behavior and employee voice: is the door really open?. Academy of Management Journal, 50(4), 869-884.
Detert, J. R., Burris, E. R., Harrison, D. A., & Martin, S. R. (2013). Voice Flows to and around Leaders Understanding When Units Are Helped or Hurt by Employee Voice. Administrative Science Quarterly, 58(4), 624-668.
Dietrich, A. (2004). The cognitive neuroscience of creativity. Psychonomic bulletin &
review, 11(6), 1011-1026.
Evans, J. S. B. (2008). Dual-processing accounts of reasoning, judgment, and social cognition. Annu. Rev. Psychol., 59, 255-278.
Grant, A. M. (2013). Rocking the boat but keeping it steady: The role of emotion regulation in employee voice. Academy of Management Journal, 56(6), 1703-1723.
Grant, A. M., & Ashford, S. J. (2008). The dynamics of proactivity at work. Research in organizational behavior, 28, 3-34.
Hayes, A. F. (2008). Introduction to mediation, moderation, and conditional process
analysis: A regression-based approach. Guilford Press.
Kane, M. J., & Engle, R. W. (2002). The role of prefrontal cortex in working-memory capacity, executive attention, and general fluid intelligence: An individual-differences perspective. Psychonomic bulletin & review, 9(4), 637-671.
Kane, M. J., & Engle, R. W. (2003). Working-memory capacity and the control of attention: the contributions of goal neglect, response competition, and task set to Stroop interference. Journal of experimental psychology: General, 132(1), 47.
Lejuez, C. W., Read, J. P., Kahler, C. W., Richards, J. B., Ramsey, S. E., Stuart, G. L. & Brown, R. A. (2002). Evaluation of a behavioral measure of risk
taking: the Balloon Analogue Risk Task (BART). Journal of Experimental
Psychology: Applied, 8(2), 75.
Liu, W., Zhu, R., & Yang, Y. (2010). I warn you because I like you: Voice behavior, employee identifications, and transformational leadership. The Leadership Quarterly, 21(1), 189-202.
MacKenzie, S. B., Podsakoff, P. M., & Podsakoff, N. P. (2011). Challenging-oriented organizational citizenship behaviors and organizational effectiveness: Do challenge-oriented behaviors really have an impact on the organization’s bottom line?. Personnel Psychology, 64(3), 559-592.
Milliken, F. J., Morrison, E. W., & Hewlin, P. F. (2003). An Exploratory Study of Employee Silence: Issues that Employees Don’t Communicate Upward and Why*. Journal of management studies, 40(6), 1453-1476.
Morrison, E. W. (2011). Employee voice behavior: Integration and directions for future research. The Academy of Management Annals, 5(1), 373-412.
Morrison, E. W. (2014). Employee Voice and Silence. Annual Review of
Organizational Psychology and Organizational Behavior, 1, 173-197.
Nijstad, B. A., De Dreu, C. K., Rietzschel, E. F., & Baas, M. (2010). The dual pathway to creativity model: Creative ideation as a function of flexibility and persistence. European Review of Social Psychology, 21(1), 34-77.
Oberauer, K., & Bialkova, S. (2009). Accessing information in working memory: Can the focus of attention grasp two elements at the same time?. Journal of
Experimental Psychology: General, 138(1), 64.
Oberauer, K., Süß, H. M., Wilhelm, O., & Wittman, W. W. (2003). The multiple faces of working memory: Storage, processing, supervision, and coordination.
Intelligence, 31(2), 167-193.
Smith, N. (2012, January 19). If You Listen Up, Your Employees Will Step Up, When you don't listen you get passive employees. Retrieved April 10, 2014, from Business News Daily: http://www.businessnewsdaily.com/1934-leadership-listening-employee-input-initiative.html
Süß, H. M., Oberauer, K., Wittmann, W. W., Wilhelm, O., & Schulze, R. (2002). Working-memory capacity explains reasoning ability—and a little bit more.
Intelligence, 30(3), 261-288.
Tangirala, S., & Ramanujam, R. (2008). Exploring nonlinearity in employee voice: The effects of personal control and organizational identification. Academy of
Management Journal, 51(6), 1189-1203.
Van Dyne, L., & LePine, J. A. (1998). Helping and voice extra-role behaviors: Evidence of construct and predictive validity. Academy of Management Journal, 41(1), 108-119.
Appendix A Demographics and Commitment Questionnaire 1. I am a
Man Woman 2. What is your age?
(Open question)
3. Are you a fulltime student at the Faculty of Economics and Business? Yes
No
4. How many points have you earned at the Faculty of Economics and Business? Less than 30 points
Between 30 and 70 points More than 70 points
5. Are you planning on following more courses at the Faculty of Economics and Business in the future?
Yes No
6. How long are you studying at the Faculty of Economics and Business? 1 year
2 years 3 years
4 or more years
7. To what extant do you feel involved in the Faculty of Economics and Business? 1-5 Likert: 1 = in a very slight degree, 5 = in a very strong degree.
