Prohibitive Voice as a Last Resort: How Doing What You Love and the Freedom of Doing Relate to Promotive and Prohibitive Voice

Prohibitive Voice as a Last Resort: How Doing What You Love and the Freedom of Doing Relate to Promotive and Prohibitive Voice

Marloes de Nekker (s2666197) University of Groningen

Department of Human Resource Management and Organizational Behavior Nettelbosje 2, 9747 AE Groningen, The Netherlands

E-mail: m.de.nekker.1@student.rug.nl

Author Note

The present paper is my master’s thesis and is written under the supervision of L. Maxim Laurijssen. Correspondence concerning this thesis should be addressed to Marloes de Nekker,

m.de.nekker.1@student.rug.nl

Abstract

Even though voice is related to important employee and organizational outcomes – including job satisfaction and performance – it is not clearly understood when and why voice emerges. This research connects two forms of autonomy – doing what aligns with enduring interests and values (concordance), and having choice (volition) – to promotive voice (improving the organization) and prohibitive voice (complaining about organizational practices), and shows how job satisfaction leads to voice rather than the other way around. It is theorized that employees first try to restore their low concordance directly, and when they fail to do so, prohibitive voice is a last resort to improve it via others. Qualitative Study 1 (N

= 108) showed that concordance and volition are distinct constructs. In line with predictions, Study 2 (N = 303; experiment) showed that prohibitive voice is most likely to emerge when people lack the volition to change their current non-concordant situation themselves.

Keywords: volition, concordance, job satisfaction, promotive voice, prohibitive voice

Prohibitive Voice as a Last Resort: How Doing What You Love and the Freedom of Doing Relate to Promotive and Prohibitive Voice

Employees can voice their opinion in the organization in different ways. They can use promotive voice with the goal of improving the organization (Detert & Burris, 2007), and prohibitive voice to express concerns (Liang, Fahr & Fahr, 2012). Research on the effects of voice is mixed, particularly for prohibitive voice. For example, both types of voice generally relate positively to job satisfaction – how people feel about their jobs and different aspects of their jobs (Spector, 1997) – suggesting that employees voice concerns to improve their organization. Other studies, however, showed that job satisfaction is negatively related to prohibitive voice, which suggests that employees voice concerns when they are unsatisfied with their current work situation (Chamberlin, Newton & Lepine, 2017). Chamberlin and colleagues (2017) argued that these inconsistencies are largely due to the lack of research that specifically focuses on how promotive and prohibitive voice differ from each other.

The present research posits that voice is largely a function of job satisfaction, and that

prohibitive voice is likely to emerge when employees are dissatisfied with the goals and tasks

they are currently working on. Job satisfaction is viewed as an antecedent rather than an

outcome of voice. It is argued that employees engage in more promotive voice when they

experience higher levels of job satisfaction, whereas employees increasingly engage in

prohibitive voice to the extent that their job satisfaction declines. In that sense, employees

may view prohibitive voice as a compensatory means to increasing their job satisfaction

(Landau, Kay, & Whitson, 2015). Particularly, the present research connects autonomy to

voice via job satisfaction and argues that employees display prohibitive voice as a last,

indirect resort to improve their work situation when they do not have the choices available to

improve their work situation directly. In order to demonstrate this, autonomy is divided into

concordance and volition.

Employees who are working in jobs in line with enduring personal values and interests experience high levels of concordance (Sheldon & Elliot, 1999), and they generally do not feel a desire to change aspects of their jobs, because their level of job satisfaction is already high. Consequently, these employees are particularly likely to display promotive voice to express their content with their work. In contrast, employees who experience low levels of concordance in their jobs do not work in line with personal values and interests. As a result, these employees are more dissatisfied, and will try to change their work environment (Radel, Pelletier, Sarrazin, & Milyavskaya, 2011). Employees’ ability to change their work situation is contingent on the availability of choices and alternative courses of action to do so, which can be referred to as volition (Ryan & Deci, 2000). Employees with jobs that are low in concordance will initially try to change their work environment directly via the choices and options they have in their job (Radel et al., 2011). Employees will turn to alternative ways to improve their situation when they are unable to do so directly due to a low level of volition (e.g., Landau et al., 2015). This research posits that prohibitive voice may be one of those alternative ways employees turn to as a last resort to have others improve their work situation when employees could not do so directly themselves.

In an ideal world, work is enjoyable and reflects employees’ enduring interests and

values. Yet, work cannot always be interesting, and boring and tedious tasks are ultimately

part of every employee’s job. Organizations may improve their functioning by providing

employees volition as a potential buffer to help them compensate for less concordant work

situations. Providing employees the opportunity to alter their routines or work methods may

help them getting through genuinely uninteresting and unvalued parts of their work. It may

also help the organization reduce the occurrence of prohibitive voice, which can be useful,

because then the negative consequences of prohibitive voice, such as negative social

consequences and late attention to problems (Liang et al., 2012), can be prevented.

Promotive and Prohibitive Voice

Employee voice generally refers to speaking up to those who have the authority to act in order to improve organizational functioning (Detert & Burris, 2007). Research found that voice relates positively to important employee outcomes, such as increased employee performance (Van Dyne & LePine, 1998), and employee engagement (Rees, Alfes &

Gatenby, 2013). These employee outcomes are also important for organizations, because they relate to increased organizational performance (e.g. Katzer, Barrett & Parker, 1961; Bakker

& Bal, 2010). However, a recent meta-analysis uncovered that there are many inconsistencies about the function of voice (Chamberlin et al., 2017). For example, some researchers have found job satisfaction as an outcome of voice, while some view it as an antecedent of voice (Chamberlin et al., 2017). Moreover, it is unclear which factors predict the emergence of voice (Chamberlin et al., 2017). Chamberlin and colleagues (2017) argue that many inconsistencies may be due to the fact that voice is a multi-faceted construct, which is often ignored. Particularly, one salient distinction in voice research is between promotive and prohibitive voice (Liang et al., 2012).

Promotive voice refers to the expression of new ideas to improve the functioning of the organization, with the goal of realizing someone’s ideals (Liang et al., 2012). The fact that employees use promotive voice in order to improve the organization makes it future- oriented (Lin & Johnson, 2015). Promotive voice is mostly used when employees want to help the organization becoming better, and when they have good intentions (Liang et al., 2012). Prohibitive voice, on the other hand, refers to the expression of concerns about problematic work practices or behaviors (Liang et al., 2012). Unsatisfactory conditions are voiced by employees (Chamberlin et al., 2017), such as an uncomfortable workplace.

Employees may also call attention to problematic practices that are misaligned with the

organization’s values (Miceli & Near, 1985). Prohibitive voice is focused on stopping or

preventing harm (Liang et al., 2012), which makes it focused on incidents that already happened in the past. This past-oriented nature of prohibitive voice seems to suggest that prohibitive voice is an outcome of job satisfaction. Employees try to solve problems they perceive (Morrison, 2011), which is oftentimes linked to negative emotions (Liang et al., 2012). Moreover, prohibitive voice is related to high interpersonal risks, because voicing dysfunction implies the failure of important stakeholders in the workplace (Liang et al., 2012). For example, when an employee has to do a lot of administrative tasks that could also be done by someone else, and he or she does not have the direct opportunity to get rid of this, this employee can complain about this to the manager by using prohibitive voice, and try to make this manager give the task to someone else or to change it in a way that makes it less time-consuming.

Chamberlin and colleagues (2017) argue that employees generally engage in voice

when they feel obligated to enact constructive change, and that employees may typically

engage in voice behavior when they have a desire to change their work situation. In the case

of promotive voice, it is clear that employees seek to further optimize their already

satisfactory work environment (e.g., Chamberlin et al., 2017). However, current literature is

mixed on the pivotal role of prohibitive voice. One function of prohibitive voice may be for

employees to change their current work situation when they are dissatisfied or unhappy

(Chamberlin et al., 2017). This seems to suggest that employees may engage in voice when

they run out of options themselves, given that voice behavior typically is displayed to those

who have the authority to act (Detert & Burris, 2007). The present research aligns with the

idea that job satisfaction maybe a precursor to voice (e.g., Chamberlin et al., 2017), and this

research develops the idea that prohibitive voice is an alternative means employees use to

change their work situation when they are unable to do so themselves.

To this end, the present research addresses autonomy as an important factor that can yield valuable insight into the function of prohibitive voice. Disentangling how autonomy relates to voice may help solving inconsistencies regards the role of prohibitive voice in the literature (e.g., Chamberlin et al., 2017). Particularly, when employees do not have the possibility to directly change their current unsatisfying work situation due to a lack of autonomy, they will search for indirect solutions to improve their situation (cf. Radel et al, 2011). Employees may turn to prohibitive voice as an ultimate way – a last resort – to try to improve this situation.

Volition and concordance

Autonomy is most widely studied within self-determination theory, which defines it as the freedom people have to engage in behavior combined with an intrinsic joy in doing so (Ryan & Deci, 2000). Self-determination theory argues that autonomy is one of the most important chronic psychological needs that people have (Ryan & Deci, 2000). Indeed, research overwhelmingly showed that autonomy is an essential nutrient in employee motivation and well-being (Ryan & Deci, 2000; see also Gagné & Deci, 2005). Autonomy is essential for facilitating personal growth, constructive social development, and personal well- being (Ryan & Deci, 2000). These findings suggest that employees are generally more motivated when they experience more autonomy. Given that autonomy relates positively to job performance (e.g., Ryan & Deci, 2000), employees’ experience of autonomy is also important to the organization.

Autonomy can be distinguished into two interrelated yet distinct constructs. First,

volition refers to the extent to which employees have choice in what they are doing in their

jobs. Volition can be defined as ‘the extent to which an individual is free to engage in a

behavior’ (Hollenbeck & Klein, 1987, p. 214). Many studies showed that volition is linked to

positive outcomes. For instance, volition is negatively related to absence (Schaufeli, Bakker

& Van Rhenen, 2009), positively related to self-esteem (Schwalbe, 1985), and negatively related to turnover intention (Spector, 1986). Moreover, research on job autonomy showed that volition is a predictor of job satisfaction (e.g. Finn, 2001; Morgeson & Humphrey, 2006;

Spector, 1986). That is, when volition increases, the level of job satisfaction is also expected to increase, because employees have the possibility to choose and change what they are working on (cf. Morgeson & Humphrey, 2006). It is typically reasoned that this may be because volition increases the likelihood that employees choose those tasks that they enjoy the most, which, in turn, increases their job satisfaction.

Second, concordance refers to situations in which employees work on goals and tasks that match their enduring interests and values, and people are found to typically choose concordant goals and tasks (Sheldon & Elliot, 1999; Judge, Bono, Erez & Locke, 2005). For example, a job is concordant when an employee finds it enjoyable and personally relevant (see also, Bianco, Higgins, & Klem, 2003). Ample evidence uncovered that concordant goal pursuit relates to positive outcomes (Sheldon, 2014). For example, according to Deci and Ryan (1985), working on interesting and valued goals and tasks leads to increased well- being, because it increases the probability of goal attainment, and people are more likely to pursue goals that will make them happy. Concordance is also positively related to job performance (Judge, Bono, Thoresen & Patton, 2001), and the perceived ability to attain immediate and desired outcomes (Sheldon & Elliot, 1999).

The present research links volition to concordance and argues that the extent to which employees engage in promotive and prohibitive voice depends on how the interplay between concordance and volition affects employees’ job satisfaction (for the research model, see Figure 1). That is, job satisfaction is predicted to mediate the link between this concordance- volition interaction and employees’ displays of voice behavior.

The Moderating Role of Concordance

Research typically showed that having volition is important to people (e.g., Deci &

Ryan, 2000; Gagné & Deci, 2005), and that working on concordant goals and tasks is important to people (e.g., Sheldon & Elliot, 1999). The interaction between volition and concordance, however, is typically not considered. This is unfortunate, as the present research proposes that this interaction may be pivotal in explaining when and why prohibitive voice is most likely to be displayed by employees.

Particularly, employees experiencing a high level of concordance in their jobs are generally happy and satisfied, because they are doing tasks that they find interesting and that are in line with their values. Concordance is positively related to job satisfaction (for a review, see Sheldon, 2014). In this situation, employees generally do not feel the desire to change their jobs, because if they had choice, they probably would have chosen the same tasks as they are already working on. Consequently, having volition is not necessary in order to increase job satisfaction when employees already experience concordance, and, therefore, mostly unimportant. Thus, for employees who experience a high level of concordance in their jobs, the relationship between volition and job satisfaction is likely to be non-significant:

Hypothesis 1a: When concordance is high, there is no significant relationship between volition and job satisfaction.

When employees are satisfied under high concordance, they are expected to engage in promotive voice instead of prohibitive voice because promotive voice is used when employees feel positively towards the organization, and want to further improve it. This also suggests that voice is a consequence of job satisfaction. Therefore, it is predicted that the interaction between concordance and volition is positively related to promotive voice via job satisfaction. Thus:

Hypothesis 1b: There is a positive relationship between job satisfaction and

promotive voice.

Hypothesis 1c: Concordance moderates the effect of volition on job satisfaction, and job satisfaction will mediate the effect of the interaction between volition and concordance on promotive voice when concordance is high.

It was proposed that employees who experience high levels of concordance do not

prefer a high level of volition, because they do not desire this volition to make a change in

their situation. Therefore, the relationship between volition and job satisfaction was expected

to be non-significant when concordance is high (Hypothesis 1a). However, employees who

experience a low level of concordance in their present work situation may become

dissatisfied, and may want to change their current situation into a more personally enjoyable

and relevant one by switching to tasks that are more concordant. Indeed, self-discrepancy

theory (Higgins, Klein, & Strauman, 1987) argues that disparities between desired and actual

psychological states can push people into action. Similarly, research on need deprivation

uncovered that people will try to restore their reduced autonomy when they desire to

experience more of it (Radel et al.,, 2011; Sheldon & Gunz, 2009). In a similar vein, it is

proposed that people who experience low concordance will have a desire to switch to

working on other tasks in their attempts to increase concordance. This is where volition

comes into play. Indeed, research showed that people turn to alternative means in order to

restore their needs (Landau et al., 2015; Zhang, Chen, Schlegel, 2018) and in their attempts to

attain personally valued goals (Kruglanski et al., 2002). This seems to suggest that under low

concordance people will want to exercise volition. That is, people’s ability to successfully

switch to potentially more concordant tasks depends on the alternative choices and options –

volition – that are available to them. Therefore, employees who experience low concordance

are predicted to want to change their situation and to improve their job satisfaction. Having

volition gives employees the direct opportunity to do this. Thus:

Hypothesis 2a: When concordance is low, there is a positive relationship between volition and job satisfaction.

However, when the level of concordance is low, and the level of volition is also low, employees do not have the direct possibility to change their jobs in a way that they enjoy it more. As a result, their job satisfaction is likely to decrease as employees run out of options to make a change. In such situations, these employees will search for alternative, compensatory strategies in order to restore their level of job satisfaction (cf. Landau et al., 2015). This means that when there is not a direct way of changing the situation, people will search for other solutions. Prohibitive voice can be seen as a compensatory way of improving the situation. Employees often engage in voice after evaluating the costs and benefits of speaking up (Liang et al., 2012). When both concordance and volition are low, the benefits can outweigh the costs, and employees may be so motivated to change their situation, that they do not even care about the risks of complaining anymore. In this situation, employees may try to use prohibitive voice as a compensatory means to boost concordance and job satisfaction when they do not have the ability to change their work situation directly via switching to different tasks. Taken together, people are least satisfied when they lack concordance and lack volition, and people may then turn to prohibitive voice as an alternative way, a last resort, to change their current situation into a more enjoyable and personally relevant situation. That is, people are predicted to engage in prohibitive voice to express their discontent if they cannot change their current situation directly in a situation of low concordance and low volition

Hypothesis 2b: When concordance and volition are both low, people are least

satisfied, display most prohibitive voice, and have a strongest desire to exercise volition.

Study 1

The main aim of Study 1 was to test people’s implicit theories towards concordance and volition. Specifically, the first aim was to uncover people’s self-reported experiences of concordance and volition, and to find out what kind of work situations people typically perceive as being concordant and volitional. The second aim of this study was to uncover what kind of activities and tasks are typically perceived as being highly and lowly concordant. The third aim was a first exploratory test of the interaction between concordance and volition. Particularly, the first prediction is that people do not express a desire for volition when they are experiencing a highly concordant situation, since they do not feel a desire to change. The second prediction is that people do express a desire for volition when their level of concordance is low, because people want to change their current situation into a more concordant one, which hinges on volition. The third prediction is that people will mention parts of their jobs they find concordant when they experience a high level of volition, because they had the opportunity to choose this themselves. Lastly, most people with low volition in their jobs are expected to write down parts of their jobs they do not like, since they do not have the ability to change their situation into a more concordant one. The results of this study are the basis for the experimental design of Study 2.

Method

Participants. A qualitative cross-sectional narrative study (Miller & Salkind, 2002) was conducted. Participants were recruited via the Prolific online platform and were paid 1.30 GBP in exchange for their participation. 14 participants were excluded from the analysis because they did not complete the story check question correctly, indicated that their data should not be used, or did not complete the questionnaire. The final sample size consisted of 108 participants who completed the questionnaire. Their age ranged from 20 to 67 years

1 Study 1 and 2 were part of another Master’s thesis (Jaleesa Donkers). Therefore, parts and analyses may overlap. Analyses and coding in Study 1 were done in a collaborative effort, and non-hypothesis specific analyses in Study 2 were shared.

(62.0% women, M

= 35.86, SD

= 11.31). 98.1% participants indicated that their mother tongue was English; the other 1.9% indicated another mother tongue. Participants’ highest level of completed education were Bachelor’s degree (44.4%), followed by High School degree (31.5%), Graduate or Professional Degree (16.7%), Master of Business Administration (4.6%), and Doctorate Degree (2.8%).

Procedure. Participants completed a three-part questionnaire. In the first part, participants were randomly assigned to one of the four conditions: high concordance, low concordance, high volition, or low volition. For every condition, a different work-related scenario was shown to the participants.

In the high concordance scenario, participants read:

“You are an employee in a large organization. As in many other jobs, there are multiple tasks that you work on during the week. All the available tasks are interesting to you. That is, these tasks are either fun or important to you, and they are in line with your enduring personal interests and values.”

In the low concordance scenario, participants read:

“You are an employee in a large organization. As in many other jobs, there are multiple tasks that you work on during the week. None of the available tasks are interesting to you. That is, these tasks are not fun and not important to you, and they are not in line with your enduring personal interests and values.”

In the high volition scenario, participants read:

“You are an employee in a large organization. As in many other jobs, there are multiple tasks that you work on during the week. You can decide which tasks you will do.

That is, amongst all the available tasks, you can choose yourself which tasks you do. What you choose, is what you will be working on.”

In the low volition scenario, participants read:

“You are an employee in a large organization. As in many other jobs, there are multiple tasks that you work on during the week. You cannot decide which tasks you will do.

That is, amongst all the available tasks, you cannot choose yourself which tasks you do. You are assigned tasks that you will be working on.”

After reading the scenario, participants were asked to remember and describe a work- related situation they considered similar to the scenario (Riessman, 2005), and they were asked how they experienced their self-reported work situation.

These work-related scenarios embodied our manipulation, and there were two main

predictions to test how concordance and volition interact via these self-reported stories

(Miller & Salkind, 2002). It was predicted that people would particularly mention aspects of

their self-reported situation that they enjoyed in the high concordance scenario, and not

mention aspects of volition, since concordance is about an enjoyable and personally relevant

experience, and these people are expected to generally not want to change their stated

situation. In other words, volition is expected to be not salient in the high concordance

scenario. The second prediction was that people would mention the lack of volition under low

concordance, provided that these people had a strong desire to change their stated situation

into a more concordant one, for instance by switching to another task or activity. Thus,

volition is expected to be salient in the low concordance scenario. Third, it was predicted that

people would mention aspects of their jobs they liked in the high volition condition, because

these people had the ability to change their situation into a preferable one. Thus, high

concordance is expected to be mentioned in high volition conditions. The last expectation

was that people would refer to situations they did not like in the low volition condition,

because these people could not change their situation into a more concordant one. Therefore,

concordance is expected to be salient when the level of volition is low.

In the second part, participants were asked to rate the concordance and volition of the personal work experience they described via self-report measures.

In the third and final part, participants were asked to describe three typical activities or tasks that relate to their scenario and self-reported work situation, and their experience of these activities or tasks. It was then measured how participants perceived the concordance and volition of these tasks by asking an open question. These questions were asked after each of the three tasks they were asked to describe.

In the high concordance scenario, participants were asked:

“Please describe what it is about this task or activity that was enjoyable/fun or personally important/in line with your values.”

In the low concordance scenario, participants were asked:

“Please describe what it is about this task or activity that made it not enjoyable/not fun or personally unimportant and not in line with your values.”

In the high volition scenario, participants were asked:

“Please describe what it is about this task or activity that made you feel you had a say or choice in this task or activity.”

In the low volition scenario, participants were asked:

“Please describe what it is about this task or activity that made you feel that you had no say or no choice in this task or activity.”

Measures

Concordance. Concordance was measured by using a self-constructed 6-item scale,

based on the Sheldon and Elliot’s (1999) definition of concordance (for all items, see

Appendix A). Participants indicated their answer on a 7-point Likert-type rating scale (1 =

Strongly Disagree, 7 = Strongly Agree). Their scores were averaged into a single

concordance score, and higher scores reflect higher experienced concordance (M = 4.26, SD

= 1.95, α = .96).

Volition. Volition was measured by using a self-constructed 4-item scale (for all items, see Appendix B). These items were based on Botti and Iyengar (2004). Participants responded on a 7-point Likert-type rating scale (1 = Strongly Disagree, 7 = Strongly Agree).

Participants’ scores were averaged into one volition score. Higher scores indicated higher experiences of volition (M = 3.39, SD = 1.76, α = .84).

Story check. Finally, participants were asked to indicate their scenario (high concordance, low concordance, high volition, or low volition) as a means of verifying whether participants were paying attention. Participants were asked whether their scenario was described as interesting (high concordance) or not interesting (low concordance), and whether they had choice (high volition) in their job or not (low volition).

Results

The answers of participants were analyzed by using an open coding technique (Corbin

& Strauss, 1990). In this technique, data are broken down analytically by an interpretative

process through multiple rounds of discussion with another coder, and consequently

conceptually similar words are grouped together to form subcategories. In this study, every

work-related scenario was first reduced into five main keywords. Next, per condition, the

most frequently mentioned and most important keywords that were conceptually judged to be

similar were grouped together into subcategories. Second, subcategories were made of tasks

that people typically mention under each scenario. Through multiple rounds of discussion,

the coders evaluated their level of agreement in coding and adjusted their coding in

collaboration. A summary of the results can be found in Table 1. Some excerpts that provide

insight in what people consider typical for each scenario are also shown in this table.

The first aim of Study 1 was to uncover people’s self-reported experiences of concordance and volition, and to find out what kind of work situations people typically perceive as being concordant and volitional. The analyses uncovered under high concordance participations mentioned tasks, activities, and situations that they want to invest their time and effort into, and they reflect enduring personal interests and values. Participants typically like to have contact with other people, solve challenging tasks, and help others. “I enjoyed all of these tasks and feel satisfied at the end of the shift when I have provided good care”.

Participants in the low concordance condition listed situations, activities, and tasks that they

(had to) spend time and effort into that are not interesting, not enjoyable, and not personally

important, for instance taking phone calls and writing reports. Participants typically listed

activities and tasks that are repetitive, that do not add value to work or personal goals, are

perceived to be time-consuming, and that they do not consider core part of their job. “I have

to write reports about my clients and then separate notes on a different system, it is repetitive

and dull”. Furthermore, under high volition, participants mentioned tasks, activities, and

situations in which they had the ability to make decisions as well as to select and prioritize

tasks. Participants typically mentioned that they like to have a variety of tasks from which

they can prioritize and have the ability to delegate. “I have a varying amount of tasks which

need to be completed per week, however the order in which they are completed is my

choice”. Participants in the low volition condition mentioned tasks, activities, and situations

in which they felt like they had no control on their activities, and they have to do a lot of

difficult work with fast approaching deadlines. Although several participants mentioned that

they have the ability to prioritize tasks, the work still feels compulsory. Participants typically

have to do tasks that are assigned to them. “I work in an organization where we work on

adhoc task basis already and we are assigned targets and to-do lists weekly/monthly on

review.” Interestingly, in both the high and low volition condition, few emotions were shared.

Moreover, participants had different interpretations and perceptions of volition. Under both the high and low volition conditions, a lot of participants experienced the ability to prioritize their tasks and decide on which one of these they want to work on. However, some participants experienced this as having high volition since they are able to decide what they want to work on, whereas other participants experienced this as low volition because at the end of the day all the tasks must be finished anyway. In the high volition condition, a participant said for instance: ‘There are often a lot of demands on my time. I tend to prioritize which has the highest clinical importance.’ A participant in the low volition condition said for example: ‘There is a long list of technical support tasks to undertake that come in for customers. We have to prioritize each one generally by the time they come in.’

The second aim was to find tasks and activities that people typically consider highly and lowly concordant. Overall, participants preferred tasks that are challenging, but not too difficult. They generally do not like repetitive and unnecessary tasks. Examples of concordant tasks are teaching and serving customers. Lowly concordant jobs are for instance taking phone calls and writing reports. Interestingly, the tasks listed by the participants are not considerably different per condition. In both the high and low concordance situations, similar tasks are mentioned, such as, for instance, data input. While participants in the low concordance condition mention their tasks more concretely, in the high concordance condition the focus lies mostly on the context of their tasks instead of the tasks themselves.

They mention for example being able to manage, create, or organize.

The third and final aim of Study 1 was to see how concordance and volition interact

with each other. First, within the high concordance scenario it was expected that participants

would not mention the desire for volition, because when people are already doing what they

enjoy and find important, there is no need to have the ability to make decisions that can

change the situation. Accordingly, only 20.7% (the least compared to the other three cells) of

the participants mentioned aspects of volition in this scenario. In the low concordance condition, it was expected that people would express a desire for volition, because these people want to change their current situation into a more concordant one. Indeed, 58.6% of the participants mentioned the lack of volition in the low concordance scenario (the highest compared to the other three cells). No participants explicitly mentioned that they desire the ability for more choice, yet they did say that they were ordered by their boss to work on the tasks and had no time to work on matters that they wanted to work on. Participants in the high volition condition were expected to mention parts of their job they found concordant, since they had the opportunity to choose this themselves. 24% of the participants mentioned aspects of concordance, particularly high concordance. Participants in the low volition condition were expected to write down parts of their job they did not like, since they do not have the possibility to change it. 36% of the participants wrote down aspects of (low) concordance in their answers. Moreover, a positive correlation was found between concordance and volition based on the questionnaire (r = .54, p < .01), indicating that concordance and volition are related to each other, although causal inferences cannot be made due to the nature of the design of this study. These results suggest that volition can help people to change their job into a more concordant one, replicating the general finding by Sheldon and Elliot (1999). Moreover, the results showed that aspects of volition are mentioned more in the high concordance volition than in the low concordance condition, and that volition and concordance are significantly related. Volition is thus mainly salient when concordance is low, which might indicate that more volition would benefit people in low concordant jobs in order to change their situation into a more concordant one. These results are initial evidence that support the expectation that there exists an interaction between concordance and volition.

Discussion Study 1

Study 1 provided initial empirical evidence that supports the idea that autonomy entails two distinct but related constructs, because participants perceived concordance and volition in unique ways. Participants indicated that they want to work on activities and tasks that they enjoy, that they find personally interesting and valuable, and that contribute to personal goals. The finding that contextual aspects are present in both the high and low concordance condition seems to reflect the idea that concordance has to with both the enjoyability of an activity as well as its personal relevance (cf. Bianco et al., 2003. Activities can be more or less enjoyable as well as more or less personally relevant, and activities may be enjoyable but not personally relevant and the other way around. In that sense, the task itself may be less important to some people than is the personal relevance of what the task helps a person accomplish. Volition is viewed as being able to select and prioritize tasks. The fact that ‘setting priorities’ is named in both the high and low volition condition can be explained by concordance, because enjoyable tasks may still be more or less personally relevant (e.g., Bianco et al., 2003; Higgins, Spiegel, Cesario, Hagiwara & Pittman, 2010).

Study 2

Study 1 showed that both the context and the task itself are important in the experience of concordance. Therefore, Study 2 relies on an experimental design in which concordance is manipulated in two different ways. First, concordance as enjoyability will be manipulated by varying the task itself. Second, personal relevance will be manipulated by adding the opportunity to leave the experiment earlier to one of the available tasks, since this is expected to be personally relevant for participants in the context of an experimental session in the research lab. Moreover, it was found that volition is perceived as having the ability to choose amongst options. Therefore, in Study 2, volition will be manipulated by giving or not giving participants the option to choose the task themselves versus being assigned to a task.

Study 2 will provide a causal test of the prediction that the interaction between concordance

and volition influences people’s displays of voice via job satisfaction, and that prohibitive voice is most likely to emerge when concordance and volition are low.

Method

Participants. An experimental study was conducted. 19 participants were excluded from the analysis, because of interruptions, software issues, influences by earlier participants, not understanding the experiment, or blatantly ignoring task instructions. This resulted in a final sample size of 303 participants who completed the experiment, ranging in age from 17 to 34 (56.4% women, M

= 20.92, SD

= 2.91). The experiment took place at the research lab at the University of Groningen, and most participants are business students. 49.8% of the participants participated in exchange for a payment of 8 euros, and 50.2% participated in order to receive research points for courses. 53.1% of the participants took the English version of the experiment, and 46.9% took the Dutch version. Most of the participants stated that Dutch was their mother tongue (45.2%), followed by German (8.3%), and English (4%).

The remaining 42.6% had a mother tongue other than these three.

Design and Tasks. For examples of the used tasks, see Appendix C. Participants were

randomly assigned to either the low or high volition condition, and it was random whether the

reference correction task or the puzzle game had a bonus attached (concordance as personal

relevance). In the low volition condition, participants were randomly assigned to the

reference correction task or to the puzzle game (concordance as enjoyability). Thus, half of

the participants engaged in the expected fun and enjoyable task (puzzle game) that did have

the bonus (high expected personal relevance) or did not have the bonus (low expected

personal relevance). The other half of participants within the low volition condition engaged

in the expected boring and annoying task (reference correction task) that did have the bonus

(high expected personal relevance) or did not have the bonus (low expected personal

relevance). Participants were not assigned to the neutral task within the low volition

condition. This task was merely present to strengthen the effect of the volition manipulation (e.g., Feldman, Baumeister & Wong, 2014). Participants in the high volition condition were free to choose any of the four tasks. Similar to the low volition condition, whether the reference correction task or the puzzle game was made personally relevant by attaching the opportunity to shorten the session by half (bonus) was varied randomly. Thus, participants could decide for themselves whether they wanted to have the bonus or not, and whether they wanted an – expectedly – enjoyable task (puzzle game) or not (reference correction task), or one of the two neutral tasks. The two neutral tasks are named differently in the experiment (“A” and “B”), but are actually the same task without the knowledge of participants, since the present research is not interested in their different outcomes. The names of the tasks the participants see on the screen are not made very concrete in order to control for the fact that participants might have done the tasks before, and make a decision based on that.

Across all tasks, participants were instructed that there were ten rounds. Participants who had the bonus task were instructed that they could shorten the task to five rounds, reducing the time spent on the task by half. Participants could shorten the task by one round by solving one round of the task correctly. Across all conditions, participants could only continue after three minutes had passed, whether they solved the current round correctly or not. This was done to ensure that time mattered to participants. After each round, participants received performance feedback. Specifically, participants saw whether they solved the current round correctly (presented in green), or not (in red), and how many rounds there were left. Participants who had the bonus task were shown by how many rounds they reduced the task, and how much time they have saved thus far.

In the reference correction task, participants saw seven reference style rules on the

screen, and in each round they had to indicate which rules were violated in each of the eight

references. In the puzzle game (Sudoku puzzle), missing numbers had to be filled in open

boxes. In the neutral task (sorting words, categorizing text), participants had to order 15 words per round in alphabetical order. Both neutral tasks were thus the same task, although participants did not know this. Both tasks were seemingly different to participants, however, both tasks were probably perceived as being linguistic.

Procedure. Before starting the experiment, all participants read and signed an informed consent form on physical paper, in which it was stressed that the experiment would last between 40 and 60 minutes. This was done to ensure that participants believed that the bonus (the possibility to leave earlier) in the experiment is actually true and not deceptive.

Thereafter, the experimenter guided the participants to one of the eight cubicles with a computer in the research lab, and started the experiment on the screen. Everything in the session was presented on the screen, except for task instructions. The experiment started with some general questions, which included demographics questions (age, sex, and mother tongue), as well as the potential theoretical covariates. Participants were then instructed that they were going to engage in an activity. On the next screen, participants were randomly assigned to a high or low volition condition. In the low volition condition, participants read:

You will be randomly assigned to one of these activities.

The experiment leader will throw a dice for you later.

The outcome of this dice will determine which activity you will perform.

Click on See Activities in order to continue.

In the high volition condition, participants read:

You will be able to choose one of these activities.

The activity you choose is the one you are going to perform.

Click on See Activities in order to continue.

Then the four tasks were shown to the participants. These are called Puzzle Game,

Sorting Words, Categorizing Text, and Reference Task. The task with the bonus was

highlighted in green and had a badge next to it that read “up to 15 min shorter”. The other tasks were not highlighted, presented in dark gray, and had a badge next to it that read “30 min”. In addition to that, participants read:

Normal (30 min) activities consist of 10 rounds and will take 30 minutes total.

However, one of these activities allow you to finish up to 15 minutes earlier.

You will shorten that particular activity by 3 min (1 round) when you solve the round correctly.

Participants were then told that they needed to go to the next screen to see which task they would actually do. In the experiment, the word ‘activity’ was used instead of task, because research has shown that the word ‘task’ is viewed as boring, which can influence concordance perceptions of participants (cf. Higgins et al., 2010). In the low volition condition, participants were told that they would be randomly assigned to a task and that they had to go see the experiment leader to throw a dice for them. The experiment leader typed the outcome of the dice on the computer screen. In the high volition condition, participants were provided the opportunity to choose any of the four tasks. On the next screen, their assigned or self-chosen task was visually confirmed. Again, the bonus task was highlighted in green, and the task that they actually would do was highlighted in blue with a large red arrow in front of it. This was done to ensure that participants know (a) which task they would do, (b) whether it was the personally relevant (bonus) task, and (c) to make visually salient whether they got the enjoyable task or not. Subsequently, participants were instructed to ask the experiment leader for instructions. These instructions were on paper and participants could see them all the time during the task. The instructions also included an example with correct answers.

Then the actual task started. After completing the task, some questions were asked. Directly

after the task, participants completed the manipulation check and dependent variable

measures. Finally, participants were thanked, debriefed upon request, and paid if they participated in exchange for money.

Volition manipulation. Participants in the high volition condition were able to freely choose one of the four activities. In the low volition condition, participants were assigned to either the high concordance task (puzzle game) or the low concordance task (reference correction task). Participants in the low volition condition were required to visit the experimenter, who would then throw a dice roll, which determined which task the participant would do. This was done to ensure that participants believed that the assignment was random and to exclude the possibility that participants believed that some of their answers causally influenced the task they would do (Wegner & Wheatley, 1999). In actuality, the software randomly assigned participants. Feldman and colleagues (2014) showed that people experience significantly more volition when there are four or more potential options to pick from, and significantly less when there are less than four options. Therefore, a total of four tasks was shown in order to assure a sufficient level of volition.

Concordance manipulation. As evidenced by previous research (Bianco et al., 2003)

and Study 1, concordance may refer to activities that are enjoyable and/or personally

relevant. Therefore, concordance was manipulated in two ways. First, concordance was

manipulated via expected task enjoyment. Similar to previous research (Isen & Reeve, 2005),

it was expected that a puzzle game is more fun and enjoyable for students than a more boring

and abstract task. Study 1 also showed that people prefer to do tasks that are challenging, but

not too difficult. Consequently, a relatively easy level of a Sudoku puzzle comprised the

highly concordant task. These puzzles were pre-tested by a convenience sample of about 10-

15 students that were approached by e-mail, Whatsapp, or physically at the faculty to make

these puzzles. This was done to ensure that these puzzles were not too simple nor too

difficult, and to ensure that puzzles could be finished in about 3 minutes. Study 1 also

indicated that a less enjoyable and uninteresting task is administrative, and a reference correction task comprised the lowly concordant task (cf. Laurijssen, Van Der Vegt, Nijstad &

Leander, 2018). There were also two neutral word categorization tasks labelled differently (“Categorizing text” and “Sorting words”), which were expected to fall in the middle of the puzzle and reference correction task (cf. Higgins et al., 2010). In addition to task enjoyment, concordance was manipulated via expected personal relevance by attaching a bonus to either the puzzle game or the reference correction task. Participants who engaged in the bonus task were able to shorten their session by half. It was expected that participants find it personally relevant and important to be as short in the lab as possible, indicated by the accumulation of experience and feedback from participants in the lab where this experiment was conducted.

Moreover, most students value their time since they may perceive their academic career as stressful (Swick, 1987), and might benefit from having more time available to work on other matters (Macan, Shahani, Dipboye & Philips, 1990), which makes it personally relevant for them to have a shorter session.

Measures

Manipulation check. Concordance was divided into two different aspects. Both concordance aspects were measured by using two adjectives that were adapted from Bianco and colleagues (2003). Every item began with the stem “I found the activity” and were followed by an adjective. For both measures, participants indicated their answer on a 7-point Likert-type rating scale (1 = Strongly Disagree, 7 = Strongly Agree). For concordance as

‘enjoyable and fun’, the two adjectives were “fun” and “enjoyable”. These were combined into one concordance as ‘enjoyable and fun’ score, with higher scores indicating more enjoyment and fun (M = 3.94, SD = 1.74, α = .93). Concordance as ‘personally relevant and important’ was measured in the same way. The two adjectives used here were “important”

and “personally relevant”. These two items were averaged into a task importance score, with

higher scores indicating higher personal importance of the task (M = 3.01, SD = 1.53, α = .78).

Volition was measured using a 4-item self-made scale based on Feldman and colleagues’ (2014) conceptualization of choice (for all items, see Appendix D). Participants responded on the same rating scale as with concordance, their scores were averaged into a single volition check score, and higher scores reflected more perceived volition (M = 3.69 , SD = 2.22, α = .91).

Task Satisfaction. Typically, job satisfaction measures include more overarching, general, and time non-specific items which try to tap into whether an employee “typically likes” the job (e.g., “I find real enjoyment in my job.”, “Most days I am enthusiastic about my job.”, cf. Bolino & Turnley, 2005). In order to conceptually measure job satisfaction within the context of a session in an experimental study in the lab which lasts maximally an hour instead of days, weeks, or years, participants were asked whether they found the task enjoyable, fun, annoying, and boring, and were asked specifically about the “activity they engaged in” instead of their job. These four adjectives were asked in the similar form as the task importance measure, and the four adjectives were adapted from Bianco and colleagues (2003) (for all items, see Appendix E). These four items together – with the latter two items reversed – were averaged into a single composite score, in which higher scores reflect more satisfaction (M = 3.99, SD = 1.67, α = .92).

Voice. Voice was measured with a qualitative question. The question was: ‘Write a review for other students who possibly want to do this session as well. How did you experience this session? Did you like it or not? Would you do something differently? If so, what would you want to change?’ In voice literature, usually voice is split into two measures:

promotive voice and prohibitive voice. Similarly, participants’ responses were coded into two

categories, based on the definitions of Liang and colleagues (2012). A participant’s response

was classified as promotive voice whenever it contained positive feedback, and text or phrases that suggest further improving the study or research. A participant’s response was classified as prohibitive voice whenever such phrases were absent and the responses solely consisted of complaining. First, participants responses were classified into a continuous measure on a self-constructed 5-point scale (1 = negative emotions and/or voice problematic practices, 3 = both types or neutral, 5 = positive emotions and/or suggestions for improvements) (M = 3.26, SD = 1.48). These scores were then also reduced into the aforementioned two categories, where numbers 1 to 3 were coded into prohibitive voice, and 4 and 5 were seen as promotive voice. (52.8% promotive voice and 47.2% prohibitive voice over all conditions). In the remainder of this text, it is explicitly mentioned whether this more continuous measure or the eventual two-category measure is used for analyses.

Desire for volition. The desire for volition participants had after performing the task was also measured. The self-constructed question, based on the definitions of Feldman and colleagues (2014), was constructed. The question was: ‘In retrospect, would you have wanted to do another task?’ Participants could answer ‘yes’ or ‘no’. Over all conditions, 34% of the participants indicated that they would have wanted to have done another task, 66% answered

‘no’.

Theoretical covariates. There were several theoretical covariates that were measured to assess their potential influence on the main analyses. Participants responded on a 7-point Likert-type rating scale (1 = Strongly disagree, 7 = Strongly agree). One item measured perceived task difficulty (“I found the activity difficult.”; M = 3.02, SD = 1.95). Participants’

perceived competence was measured with three items (“I felt competent in this activity”, “I

felt able to master this activity”, and “I was good at doing this activity.”; M = 5.12, SD =

1.64, α = .90). Participants’ experience of time pressure was measured with three items (“I

felt time pressure”, “I felt like having too little time”, and “I would have liked to have more time”; M = 2.29, SD = 1.37, α = .83).

Results

Preliminary analysis. A multivariate ANOVA with the three theoretical covariates

perceived difficulty, F(2, 300) = 63.42, P < .0001, perceived competence, F(2, 300) = 56.16, P < .0001, and perceived time pressure, F(2, 300) = 14.13, P < .0001, as dependent variables and the actual task participants did (reference correction task, neutral task, or puzzle game) as independent variables revealed significant effects of these theoretical covariates. Similarly, a regression analysis revealed that perceived time pressure, b = 0.23, SE

= .06, t(299) = 4.11, p

< .0001, and perceived competence, b = .55, SE

= .07, t(299) = 7.80, p < .0001, but not perceived difficulty, b = -.09, SE

= .06, t(299) = -1.43, p = .16, were significantly related to task satisfaction. Perceived competence and perceived difficulty share a large conceptual overlap and may have reinforced each other, r = -.76, p = .01. Participants felt most competent in the puzzle game (M = 5.90, SD = 1.03), somewhat less in the reference correction task (M = 4.47, SD = 1.73, and the least in the neutral task (M = 3.70, SD = 1.71).

Participants perceived the neutral task to be the most difficult (M = 4.85, SD = 1.77), followed by the reference correction task (M = 3.80, SD = 1.98), and the puzzle game (M = 2.06, SD = 1.32) was comparatively perceived to be the easiest. Finally, participants perceived most time pressure in the neutral task (M = 2.71, SD = 1.69), followed by the puzzle game (M = 2.55, SD = 1.31), and the reference correction task (M = 1.74, SD = 1.16).

Taken together, in the main analyses, there was controlled for perceived time pressure, perceived competence, and perceived difficulty, unless explicitly stated otherwise.

Manipulation check. Both the concordance and volition manipulation were built

around two focal tasks that were expected to be more (puzzle game) or less (reference

correction task) concordant. In order to see whether these expectations were confirmed by the

data, two ANOVAs were run, each with the task (reference correction task, neutral task and puzzle game) as the independent variable.

The first ANOVA revealed that indeed the puzzle game is perceived to be more fun and enjoyable (M = 4.76, SD = 1.43) and the reference correction task (M = 3.10, SD = 1.58), and the neutral task was a bit less enjoyable than the reference correction task than the neutral task (M = 2.82, SD = 1.63), F(2, 300) = 51.04, p < .0001. The second ANOVA revealed that the reference correction task is perceived to be more personally relevant and important (M = 3.46, SD = 1.71) than the puzzle game (M = 2.89, SD = 1.37), and the neutral task is least personally relevant and important (M = 2.28, SD = 1.27), F(2, 300) = 10.05, p < .0001. Thus, the puzzle task is perceived to be more fun and enjoyable than it is personally important and relevant, whereas the reference correction task is generally perceived to be more personally important and relevant than it is fun and enjoyable.

Next, two ANOVAs were run to find out whether the three manipulations (concordance as fun and enjoyability via the actual task, concordance as personal relevance and importance via the bonus, and volition) worked. In the first ANOVA, the three categorical variables (concordance as fun: reference correction task, neutral task, puzzle game; concordance as personal relevance: low vs. high; volition: low vs. high) were entered as independent variables and the concordance manipulation check was the dependent variable. Only the actual task participants engaged in – concordance as enjoyability – significantly influenced participants’ overall perceptions of concordance, F(2, 294) = 17.68, p

< .0001, not whether the task had the bonus (concordance as personal relevance), F(1, 294) = 0.44, p = .51, nor the volition manipulation, F(1, 294) = 2.33, p = .13. There were no significant interactions (Ps between .20 and .46). On average, participants rated the puzzle game as most concordant (M = 3.73, SD = 1.08), followed by the reference correction task (M

= 2.94, SD = 0.97) and the neutral task (M = 2.81, SD = 0.95). There were also measures that

directly tapped into whether the task was enjoyable and fun or whether it was personally relevant and important after the task. This revealed that only the actual task (concordance as enjoyability) significantly influenced the self-reported enjoyment and fun of the task F(2, 294) = 36.43, p < .0001, not the other manipulations or their interactions (Ps between .18 and .50). The manipulations and their interactions did not significantly influence participants’

experienced personal importance and relevance of the task (Ps between .42 and .96).

The second ANOVA was identical to the first one, but this time the volition check was the dependent variable. This analysis revealed that participants experienced significantly more volition in the high volition condition (M = 5.95, SD = 1.28) compared to the low volition condition (M = 2.27, SD = 1.32), F(1, 294) = 50.41, p < .0001. There was also a significant main effect for whether participants had the bonus task or not (concordance as personal relevance), F(1, 294) = 19.73, p < .0001, as well as for the actual task participants engaged in (concordance as enjoyability and fun), F(2, 294) = 7.97, p < .001, and all interactions were significant (Ps < .01). The effect of the actual task is likely a statistical artefact of the design, and driven by the neutral task. Participants could only choose the neutral task in the high volition condition, and the task was absent in the low volition condition. Unsurprisingly, therefore, the mean for the volition check for the neutral task (M = 6.04, SD = 1.09) is significantly higher than those for the other tasks that were present in both the high and low volition condition (reference correction task M = 2.45, SD = 1.66; puzzle game M = 3.94, SD = 2.20).

Taken together, the concordance as enjoyability/fun and volition manipulations have

worked, whereas the concordance as personal relevance/importance manipulation has not

worked. Whether the bonus was present or not did not influence the perceived

relevance/importance of the task, though the reference correction task was perceived to be

significantly more personally relevant/important than the puzzle game, regardless of the manipulations.

Hypothesis Testing

For all hypotheses testing, participants that performed the neutral task were excluded for analyses because it was not part of the concordance manipulation, and including the neutral task would lead to an unbalanced design since the neutral task is only present under high volition, not low volition As a result, the remaining main analyses were performed with N = 264 participants.

Hypothesis 1a and 2a. The first predictions state that concordance is linked to satisfaction, and that concordance moderates the link between volition and satisfaction such that volition is non-significantly related to satisfaction under high concordance (Hypothesis 1a) but significantly positively related to satisfaction under low concordance (Hypothesis 2a).

In order to test this, an ANOVA was run with the task participants engaged in as the independent variable (concordance as enjoyability and fun), as well as whether they had the bonus task (concordance as personal relevance and importance) and self-reported satisfaction as the dependent variable. This analysis revealed that participants indeed experienced more satisfaction when they engaged in the highly enjoyable task (puzzle game, M = 4.77, SD = 1.39) than when they engaged in the lowly enjoyable task (reference correction task, M = 3.21, SD = 1.52), F(1, 257) = 12.25, p = .001. Participants also experienced more satisfaction when they had the task that was made personally relevant by having the bonus (M = 4.32, SD

= 1.66) compared to not having the bonus (M = 3.93, SD = 1.57), F(1, 257) = 11.08, p = .001.

There was no significant interaction between the actual task and whether participants had the

bonus task, F(1, 257) = 0.94, p = .33. In short, participants experienced more satisfaction

when the task was highly concordant than when it was lowly concordant, both in terms of