The indirect effect of voicing workers on organizational success : assessing the influence of an employee’s voice on the beneficial behavior of co-workers

(1)

The indirect effect of voicing workers on

organizational success

Assessing the influence of an employee’s voice on the

beneficial behavior of co-workers

Amsterdam, June 30, 2014 Student: Lucia Pot, 6069541

Thesis seminar Business studies Coordinator: Hannah Berkers Supervisor: Renske van Geffen Academic year 2013-2014 Semester 2, Block 3

(2)

Abstract

The voice of employees being a beneficial influence on an organization is a thoroughly studied subject. Yet, the fact that a worker’s voice might affect the Organizational

Citizenship Behaviour of his or her co-workers and thus indirectly affect the organization, is a phenomenon that is not covered in the literature. To explore a relationship between an employee’s voice and a co-worker’s subsequent behaviour, individuals’ reactions to certain voice-contents are analyzed. We propose that acts of voice that differ in their message will cause nuances in resulting behaviour. More specifically, we suggest that promotive voice, prohibitive voice and suggestion voice will cause varying behavioural reactions within

general OCB, challenging OCB and affiliative OCB. In addition we propose that tenure of the employee who receives the voice, will have a moderating effect on these relationships since this factor can affect the way a person reacts in advance. We test these hypotheses by the use of a diary study lasting a week. A valid sample of 272 respondents is used, all being fulltime workers or full time interns operating in Dutch companies. The propositions considering the direct effects between voice forms and OCB, are partially supported. This suggests that the voice of co-workers influences the behaviour of employees and that details like varying messages do matter in such research. The interaction effects, designed to analyze the

moderating effect of tenure, are not supported. The amount of years a co-worker has worked for an organization does not seem to influence the relationship between an employee’s voice and that co-worker’s subsequent behaviour. .

(3)

Foreword

This thesis on the act of voice and its influence on behavior within an organization, was written for my Bachelor degree in Business Studies at the University of Amsterdam. I would like to thank the people that helped me complete the research, making it possible for me to write this thesis. Special thanks is to be given to my supervisor, Renske van Geffen, for her help and vision regarding the plan of analysis and the regular feedback. Also, I would like to thank all that participated in the diary study and thus dedicated a week of their attention to filling out the surveys.

Then all that is left to say is that I hope you will enjoy reading my thesis.

Lucia Pot

(6)

1. Introduction

The modern economy is changing dramatically. From being industrial and locally focused in the past, the economy is now global and based on knowledge (Morgan, Campbell, Crouch, Pedersen & Whitley, 2010). This increases the intensification of multimarket competition and thus of rivalry (Van Witteloostuijn, 1993). To stay at the top, it is essential for an

organization to create inimitable competitive advantages (Reed & DeFillippi, 1990). There are numerous ways to achieve such advantages, yet the workforce of an organization is especially seen as a valuable aspect in this matter. It is crucial to accomplish beneficial organizational behavior amongst employees since they are important within an organizational climate and therefore substantially responsible for driving success (Brondino, Silva, & Pasini, 2012).

A construct of beneficial behavior amongst employees is Organizational Citizenship

Behavior (OCB). In simple terms, it can be seen as a global concept that includes all positive

organizational behaviors of individual organization members (Graham, 1991). When realized, it can improve a firm’s sustainability, performance, long term growth and profitability

(Chowdhury, 2010). Successful OCB among personnel is therefore a valuable competitive advantage for companies to strive for.

It is important for organizations to understand how to achieve a competitive

advantage like OCB among personnel. The popularity of modern-day employee involvement

programs displays this importance (smallbusiness.chron.com), since involvement is proven to

be a substantial predictor of OCB (Diefendorff, Brown, Kamin, & Lord, 2002). A company that successfully involves its employees and turns them into a competitive advantage is furniture manufacturer Ikea. Amongst other things, Ikea is renowned for its workforce being one of its main strengths. It makes use of shared values among workers, which produces positive organizational behaviors of individual organization members (Graham, 1991). These

(7)

workers stand on the company’s side with the desire to learn and improve and are therefore more inclined to get involved and show OCB. Ikea uses the act of voicing to achieve these shared values and thus OCB. They developed the Voice-survey, which shows a clear connection between voice-results, shared values and business results (Ikea, 2006). It seems that voicing amongst employees is desirable for organizations, according to LePine and Van Dyne (1998) it is a form of proactive behavior. Thus, the voice of a worker can be seen as a direct contribution to OCB. This fact is already extensively covered in the literature and considered by many companies like Ikea.

Still, little research has been done on the effect that a worker’s voice has on the behavior of his or her co-workers, the knowledge is therefore limited. As a result, the

question remains whether the voicing of an employee also indirectly affects the organization through the resulting changing behavior of co-workers. This extra dimension could help a company like Ikea take its employee policy to a new level.

A person that has been working for a company for an extensive amount of years will most likely show a different reactive behavior to a voicing colleague than a new employee, since gained personal experience has a powerful effect on how a person assesses a situations and acts upon it (Weinstein, 1989). Therefore, it is expected that a high tenure will cause different reactions to changes and suggestions.

Consequently, this study aims to find the relationship between voicing of co-workers and the resulting behavior of employees. Also, its intention is to examine the moderating effect that tenure has on this relation. This leads to the following research question:

“What effect does the voicing of a co-worker have on the Organizational Citizenship Behavior of an employee and how is this effect influenced by the tenure of this employee?”

(8)

2. Literature review

In this section, existing literature that is relevant to our research problem will be exhibited in order to create an understanding of the subject of study. A short conclusion will follow in which the need for this research will be emphasized, also a fitting research question will be provided.

2.1 Organizational Citizenship Behavior (OCB)

Achieving a competitive advantage through the beneficial behavior of workers is a complex matter. As Porter (2008) stated, a competitive position often cumulates from many activities so it is not easy to diagnose and imitate. Due to the intricacy of this matter, it is suitable to focus on one behavioral construct, namely Organizational Citizenship Behavior (OCB). This construct has received a lot of scientific attention in recent years (Huang, You, & Tsai, 2012), it appears in many different forms, from conscientiousness, to courtesy, to sportsmanship and more. One instance of OCB may not be substantial, since it can be the smallest beneficiary act for the company. Yet, all acts combined this discretionary behavior can have a key beneficial impact on operations and effectiveness within an organization (Organ, 1988). Added to that, it has often previously been confirmed with empirical data that OCB positively influences organizational success, which makes it a highly useful subject to study (Podsakoff & MacKenzie, 1997).

With a rapid growth in literature on OCB and related subjects, it is clear that there is an increasing need for knowledge on this topic (Podsakoff, MacKenzie, Paine, & Bachrach, 2000). This also has the unfortunate consequence that there are thirty different conceptual definitions of the construct of OCB, of which most do show some overlap. These can be placed under seven common dimensions: 1) Helping behavior, 2) Sportsmanship, 3) Organizational Loyalty, 4) Organizational Compliance, 5) Individual Initiative, 6) Civic

(9)

Virtue and 7) Self Development (Podasakoff et al., 2000). This division might however be too extensive to use for some researches, thus there also exist more compact definitions of OCB. An example is the division of OCB into: The affilitative form, which is intended to maintain and enforce the status quo and the challenging form of OCB, which for instance makes use of constructive changes (Choi, 2007). Hence, one must choose a fitting definition of OCB for their research.

When looking at antecedents of OCB, a large pool of variables can be studied. Yet, according to Podasakoff et al. (2000) a more complete set of variables needs to be created and investigated for this purpose. Therefore, researching the voice of co-workers in relation to OCB may provide the academic world with useful new information. In prior research the focus was mainly on the following four categories of antecedents: Individual (the employee), task characteristics, organizational characteristics and leadership behaviors (Bateman & Organ, 1983; Organ, 1988; Smith, Organ, & Near, 1983). There has been no mention of the voice of co-workers as an antecedent of OCB and it does not fit within one of the mentioned categories. Hence, there exists a literature gap when it comes to the influence of co-workers’ acts of voice on the OCB of employees.

2.2 Importance of co-workers

In this section several occasions will be presented in which co-workers tend to be an

influence through the use of their voice, showing that it is a defect not to incorporate it as an antecedent of OCB. For instance, co-workers can provide each other with useful and

educational information, here the act of voice is clearly apparent. Also, co-workers can show behavioral support for beneficial practices, this could be in the form of words of

encouragement. Other practices might be discouraged, for instance by the use of a warning. (Ashforth, 1985; Chiaburu & Harrison, 2008). Such verbal support or discouragement can cause co-workers to shape each others’ roles and consequently influence behavior (Ashforth,

(10)

1985; Chiaburu & Harrison, 2008). Co-workers also influence each other by promoting or rejecting certain norms, these norms tend to regulate human behavior (Greenberger, Miceli, & Cohen, 1987; Feldman, 1984). Thus, this promotion or rejection will consequently

influence human behavior, this can be expressed by for instance speaking up one’s opinion. Within an organization, workers can have different organizational backgrounds whereby the influence of their voice may differ. For example, there exist differences between the influence of an immediate co-worker, an external organization co-worker or a leader speaking up. Therefore, when researching the effects of a co-worker’s voice, it might be appropriate to create a framework in which diverse types of co-workers are incorporated (Greenberger et al., 1987). However, this adds complexity to the research and due to the time constraint this might be better to avoid for now.

2.3 Voice among workers

On the occasion that a worker ‘speaks up’, this is known in the literature as an employee who ‘voices’. The definition of voice behavior according to LePine and Van Dyne (1998, p.853) is the following: “speaking out and challenging the status quo with the intent of improving the situation”. For this reason, voice can be seen as a form of proactive behavior, a relatively stable tendency to effect environmental change (Bateman & Crant, 1993).

Diverse research streams state that voicing workers have significant benefits for organizations. However, it does not only directly impact surroundings, it also impacts the people who voice themselves for instance (Grant & Ashford, 2008). Yet, the fact whether or not the voice of a worker impacts the behavior of his or her co-workers remains scarcely researched and therefore it evokes our interest.

Wanting to positively change a situation stands in line with OCB, voicing of an employee can therefore occasionally be seen as an expression of OCB with a direct influence on organizational success. Yet, in this research the voice of a co-worker is looked at as an

(11)

independent act which influences the OCB of employees, thus it is treated as an antecedent of OCB. When using this viewpoint, one looks at the indirect effect a worker's voice has on the success of an organization. It is important to not confuse these interpretations and measure what is relevant to the research in order to retain validity.

Voice can be an intricate multidimensional concept and it can be categorized in several ways, for instance based on motives. This would lead to Acquiescent Voice,

Defensive Voice, and ProSocial Voice (Dyne, Ang, & Botero, 2003). However, motives can be seen as the underlying intent of a person’s actions (Ho, ForsterLee, ForsterLee, & Crofts, 2002). This makes motivational information hard to gather, therefore it is a difficult

categorization to use. Considering the time constraint, it is better to use information that lies more on the surface. Taking this into account, one can for example choose a categorization based on the different messages that an employee tries to bring across. This could for instance lead to a construct that contains promotive voice, which are expressions to improve existing work practices and procedures to benefit organizations (Liang, Farh, & Farh, 2012). This stands next to prohibitive voice, which contains expressions of individuals’ concerns about existing practices, incidents or behaviors that may harm their organization (Liang et al., 2012). Another example is the construct of Morrison (2011), in which voice is defined as a discretionary communication of ideas, suggestions, concerns or opinions about work-related issues with the intent to improve organizational functioning.

For this research, a categorization based on the differences in messages of voice is chosen since this required information lies more on the surface. Among the mentioned voice constructs of Morrison (2011) and Liang et al. (2012), overlap exists. Promotive voice can for instance be compared with communication of ideas. Likewise, prohibitive voice can be linked to the communication of concerns. Furthermore, the communication of opinion is presented in both promotive and prohibitive voice, as people state how they feel about current

(12)

related issues and change. This shows that Morrison (2011) has added an extra dimension with the communication of suggestion. In this form of voice, the person tends to offer suggestions on how a practice could be changed (Morrison, 2011).

Due to the different existing descriptions of voice, a single categorization has to be selected when conducting research. Consequently, in this research it is chosen to combine the two mentioned constructs. Suggestion voice of Morisson (2011) is added to the definition of Liang et al. (2012), as this extra dimension is believed to enhance the depth and explanatory power of our research. Thus, in our study a new voice construct of promotive voice,

prohibitive voice and suggestion voice will be analyzed.

2.4 Tenure of employees

When analyzing tenure of employees, it can be useful to firstly consider turnover. As Liu, Zhang, Wang and Lee (2011) demonstrate, the presence of autonomy and added support keeps employees in their jobs. Thus, people with a long tenure are expected to enjoy more autonomy. This form of independence will most likely cause these employees to also be more confident and have more power, as these three factors are often linked (Goldberg &

Wooldridge, 1993).

When an employee has been working for the organization for a long period of time, he or she typically gains employer specific skills ,which are usually achieved by regular practice (Ericsson, Prietula, & Cokely, 2007). It is plausible to assume that employees with high tenure tend to enjoy more employer specific skills and experience, since they have had more time to practice. These skills often come with gained stability and a tendency for workers to fall into a routine (Turner & Fern, 2012). Yet, when workers become slaves to routine they do not respond well to changing conditions (Edmondson, Bohmer, & Pisano, 2001). Workers with a high tenure therefore are more likely to be submissive to habits and previously acquired skills, this will make them reluctant to change and less driven to

(13)

positively react to new ideas and suggestions.

There are several known predictors of tenure, a strong example is the age at which a person joins a company. The younger a person is at that moment, the longer he or she tends to serve at that organization. Also, people that stay with an organization for a long time tend to remain in the same position (Cramer, 1993). These predictors show a common factor, namely commitment. Commitment to a company and commitment to a job. It indicates that people with long tenure will most likely feel more obligations towards their organization.

A positive experience of work relationships enables employees to feel secure, therefore it can influence their behavior (Vinarski-Peretz, Binyamin, & Carmeli, 2011). Tenure can have an influence on the experience of such relationships. According to Dutton and Heaphy's (2003), there are three subjective relational experiences: positive regard, mutuality and vitality. Positive regard corresponds to a feeling of being known or loved, being new in an organization will make this regard less feasible. Mutuality is a worker’s sense that both parties are actively participating in a positive relationship and it encourages individuals to engage in shared activities (Dutton & Heaphy's, 2003). The active participation in the relationship might not be equal when tenures differ since newer employees might be more needy of this. Therefore the employee might be less inclined to engage in shared activities, thus their behavior is influenced. Relational vitality refers to feelings of positive excitement and a heightened positive energy arising from one's relationships with co-workers, which makes a person driven to look forward to every new challenge (Spreitzer, Lam, & Fritz, 2010). When tenures differ, this energy is most likely one-sided and therefore will not heighten. The attitude towards a new challenge might therefore be less exited due to the calmness of workers with a high tenure.

It appears that tenure of an employee influences the way a person experiences work relationships, this causes it to also influence eventual behavior.

(14)

2.5 Conclusion

Organizational Citizenship Behavior (OCB) is a renowned construct-form of behavior among employees and it has a key beneficial impact on operations and effectiveness within an organization. Co-workers can influence each other and are therefore important for each other’s organizational behavior, thus also for eventual organizational success. The act of voice of a co-worker can therefore be analyzed as an antecedent of OCB, instead of it being a form of OCB that directly influences organizational success.

Tenure appears to be an influential factor when analyzing how an employee stands within a company and how he or she responds to certain situations, therefore this moderating factor is expected to produce interesting views.

The relationship between a co-worker’s voice and a worker’s behavior has not been studied before, especially not with the moderating factor of tenure. Yet, existing literature seems to support its plausibility and importance. This gap in the literature leads to the following research question:

“What effect does the voicing of a co-worker have on the Organizational Citizenship Behavior of an employee and how is this effect influenced by the tenure of this employee?”

(15)

3. Conceptual framework

3.1 Influence of a co-worker’s voice on (Challenging) OCB

To analyze whether the act of voicing of a co-worker will have an effect on the behavior of employees, human culture is taken into account. This is an important factor when considering the manner in which people act and react in certain situations. Evidently, it is human culture to copy others as a strategy to learn (Rendell, Boyd, Cownden, Enquist, Eriksson, Feldman, & Laland, 2010). A person who voices shows proactive behavior, thus it is expected that in order to learn, this proactive attitude will be copied. When invested in behavior, it will most likely result in more OCB.

As previously stated, co-workers tend to regulate each other’s behavior (Feldman, 1984), it is therefore expected that employees are partly shaped to the liking of co-workers. With voicing workers being proactive types, it is assumed that the behavior of colleagues is shaped to being proactive also.

Thus, when a co-worker speaks up and shows pro-active behavior, it is anticipated that the employee will also enhance his or her OCB. This leads to proposition 1:

Proposition 1a: The OCB of an employee is positively affected by the voicing of a co-worker

The affiliative and challenging form of OCB are very different and are therefore expected to produce alternative benefits for an organization (Choi, 2007). It is worth discovering which part of OCB will be aroused due to the voice of a co-worker.

Due to the human culture to copy and co-workers’ tendencies to shape each other’s roles (Rendell et al., 2010; Ashforth, 1985), it is expected that if the voicing of a co-worker has an effect on an employees’ behavior, it will mostly enhance their OCB in the challenging

(16)

form. The act of voicing challenges and possibly upsets the status quo, this relates to the

challenging form of OCB (Detert & Burris, 2007). This leads to proposition 1a:

Proposition 1b: The Challenging OCB of an employee will be more affected than the

affiliative OCB of the employee by the voicing of a co-worker

3.2 Differing effects on OCB of diverse dimensions of voice

As mentioned, the new voice construct of promotive voice, prohibitive voice and suggestion voice will be used. This firstly leads to a proposition that incorporates promotive voice, where a co-worker would for instance mention new ideas and lean towards change. Consequently, literature is used that studies the way a person reacts to change. A widely accepted mental model that drives organizational behavior is the idea that there is a general resistance to change and that participation is the primary method to overcome it (Dent & Goldberg, 1999). When co-workers voice about change, employees will be incorporated in the idea and be made to feel like participants. Accordingly, it is expected that this will greatly reduce the natural resistance towards the change. Hence, the following proposition is created:

Proposition 2a. The OCB of an employee is positively affected by the promotive voicing of a

co-worker

It is believed that promotive voice of a co-worker will lead to challenging OCB in an employee, as both focus on change. Since the participation takes away resistance, the

employee is expected to show behavior incorporating the suggested changes. This leads to the following proposition:

Proposition 2b. The challenging OCB of an employee will be more affected by promotive

voice than the affiliative OCB of an employee

(17)

As mentioned before, prohibitive voice contains a very different message than promotive voice. As a result, a different behavioral reaction of the employee is expected. Concerns with current ways can easily lead to an air of frustration, especially since

prohibitive voice does not incorporate a new plan. It will cause people to react excessively when they comprehend their own fallibility and the world's ingratitude (Times). Such organizational frustrations have a negative effect on the behavior of organizational

employees, it can cause them to waste time and materials (Spector, 1975). Consequently it is believed that prohibitive voice will have a negative effect on behavior, which brings us to the following proposition:

Proposition 3. The OCB of an employee is negatively affected by the prohibitive voicing of a

co-worker

Lastly, voice in the form of suggestion will be studied. The suggestion to change something can be interpreted as giving someone advice. Therefore, in this part the focus lies on how an employee reacts to the advice of a co-worker.

According to Bonaccio and Dalal (2006), the utilization of advice by the receiver (the judge of the advice), is influenced by the extra expertise or knowledge that an advisor

possesses in contrast to the judge. This would give the adviser expertise power and the judge would be less inclined to discount the advice. But since in this case peers are analyzed, there is no expectation of expertise power being present. The receiver will therefore tend to discount the advice and consequently not be influenced by the voice. This leads to the following proposition:

Proposition 4. The OCB of an employee is not affected by the suggestion voice of a co-

Worker

(18)

3.3 The effect of tenure on different dimensions of voice

As previously mentioned, often people do not effectively utilize advice. In fact, studies have also consistently shown a negative relationship between power and advice taking, this effect is mediated by confidence (Morrison, Rothman, & Soll, 2011). Power and confidence can thus enhance people’s tendency to prefer their own initial judgment and not react to voice, neither change their behavior. People with high tenure will in most cases experience more confidence and seemingly more power (Goldberg & Wooldridge, 1993). Therefore it is plausible that tenure of an employee will negatively influence the relationship between voicing of a co-worker in the form of a suggestion and that employee’s OCB. This leads to the following proposition:

Proposition 5a. The higher the level of tenure, the less affected an employee’s OCB is by the

suggestion voice of a co-worker

As previously mentioned, employees practicing to gain expertise can cause

submission to habits and slavery to routine. It is expected that a co-worker with high tenure will be more sensitive to this, due to the many years of experience. Consequently, these workers will be less inclined to change their behavior when a co-worker wants to alter existing work practices and procedures. Thus, this promotive voice will have less effect due to high tenure. This leads to the following proposition:

Proposition 5b. The higher the level of tenure, the less affected an employee’s OCB is by the

promotive voice of a co-worker

Due to the higher feeling of commitment towards the organization that workers with high tenure mostly experience (Cramer, 1993), he or she will tend to feel personally

responsible to deal with discretionary behavior (Morrison, 1994). This leads to the idea that a

(19)

worker with a high tenure will change his or her behavior when a problem or concern is mentioned by his or her colleague, which is the case with prohibitive voice. This leads to the following proposition:

Proposition 5c. The higher the level of tenure, the more positively affected an employee’s

general OCB is by the prohibitive voice of a co-worker

Thus, two of three forms of tenure negatively influence OCB and as previously mentioned it seems to unconstructively affect the positive regard, mutuality and vitality within work relationships. Therefore, the relationship between general voice and OCB is expected to be negatively influenced by tenure.

Proposition 5d. The higher the level of tenure, the more negatively affected an employee’s

OCB is by the voice of a co-worker

(20)

4. Methodology

4.1 Level analysis and applicability

The level of analysis that was used in this research was that of fulltime workers and fulltime interns within the Netherlands. The sector or the organization in which the participants operated made no difference, since voice is an enduring workplace phenomenon. Four in five workplaces, no mind the sector, incorporate situations that provide for voice. It takes place in all firms and organizations that hold two-way communication mechanisms used by fulltime workers or fulltime interns (Brown, 2009). Therefore, the conceptual model that resulted from this study is applicable to all organizations within differing sectors. However, since the research was based on co-workers, these organizations had to incorporate at least two

employees.

4.2 Study Design

This study was done in the form of an empirical research in which the derived knowledge was personally gained. It was based on the evidence of indirect observation by means of a diary study, subsequently it was quantitatively analysed. Use was made of a single data collection technique and corresponding analysis procedures. The diary study was in the form of self-administered surveys that were filled out three times a day for a full workweek. This was a form of experience sampling, which is a method of data collection in which

respondents participate in repeated assessments at specific moments over a course of time. Also, the participants functioned in their natural settings (Scollon, Kim-Prieto, & Diener, 2003).

The diary study was used because participants were then able to report on the

experiences and events of their daily lives. Also, due to modern technology it was minimally intrusive, which made the study highly reﬂective of the participants (Bolger, Davi, & Rafaeli,

(21)

2003).

To explain it more precisely, in this study a pre-test was firstly sent on the Saturday morning prior to the workweek in which the diary study began. This was filled out by respondents in order to collect data that did not change during the length of the diary-week. In this pre-test, questions concerning the personalities of participants and some demographics were mostly asked, like age or tenure. Unfortunately, during our research not all respondents properly received this pre-test. These participants were sent the pre-test again on the Saturday following the diary-week. Since this knowledge was unchangeable, the delay did not cause substantial implications.

Added to this, respondents were also asked to fill out three short surveys a day during a particular work-week, the diary week. They had to work all of the five days (Monday until Friday) and the times on which the surveys were sent were suited to common work-hours (09 am until 05 pm). Participants received a survey in the morning before work, a survey in the afternoon during lunch-hours and a survey in the late afternoon/evening to be filled out after work. When certain jobs varied from common working hours, the moments on which those surveys were sent were adapted to fit the situation of that respondent. After a specific amount of time, the chance to fill out a certain survey expired to prevent participants from filling out several surveys at once. This would have defeated the purpose of using a diary study. Surveys were well suited for this study, the standardized questions made it easier to compare the data of a large amount of people (Saunders, Lewis, & Thornhill, 2009).

Considering the research question, there was a need for comparisons of for instance behaviors and contents of voice. All surveys used in the diary week and the survey used for the pre-test, were computer-administered. This was less expensive, cost less time and a larger amount of people could be reached. Added to that, the transparency for the respondents was enhanced (Rosenfeld, Booth-Kewley, & Edwards, 1993). This was valuable, since there was a limited

(22)

timeframe to achieve results within this research. The self-administered surveys entailed less process-time and reduced the bias of the participant or subject that could have emerged if answers had to be written down in interviews (Saunders et al., 2009). This manner of diary study was also chosen to avoid the risk of being an intrusive and/or a reactive force among the workers, which also could have created a risk for bias (Jenkins, Nadler, Lawler, & Cammann, 1975).

An acknowledged limitation of using surveys is the fact that people do not wish to fill out a large questionnaire (Saunders et al., 2009). For this reason the surveys used in the diary week were small enough to be filled out within two minutes, as a result there was less

resistance. The pre-test was longer and took participants approximately ten minutes to fill out. Yet, it was not experienced as being inconvenient, since the pre-test was only sent once before or after the diary surveys.

The focus of this study was on Dutch fulltime workers and Dutch fulltime interns within all organizations and all sectors. A reliable sample was drawn from the mentioned population which made the results generalizable in order for the new knowledge to be of wide use (Saunders et al., 2009). Within this population simple random sampling was used in order to avoid bias (Saunders et al., 2009).

4.3 Research Sample

In this research probability sampling was used, this is most common when conducting survey-based research. In order to avoid bias, simple random sampling was chosen and it was strived to attain a sample as large as possible to make results generalizable. A minimum sample size of thirty people is required in order to generate data with a normal distribution, which makes it functional for statistical testing (Stutely, 2003, in Saunders et al., 2009, p. 218). With six students all striving to collect thirty or more participants, together a decent

(23)

sample of at least 180 respondents was expected. When all collected, the total amount of respondents was 194. This sample was large enough to remain a decent sample size whilst accounting for the possibility of non-respondents.

The diary study reflected the behaviors and acts of voice in the work environment. Repeatedly asking about these factors through a whole workweek produced an extensive representation in which the use of a chronological format helped to identify development.

4.4 Operationalization constructs and variables

The pre-test contained unchangeable variables, including data about the respondents’ characteristics and personality (Saunders et al., 2009). In the daily surveys the focus was mostly on opinion variables, to analyze how respondents felt about something or whether they believed certain facts to be false or true. It also incorporated behavioral variables, which contained data on how participants actually acted.

4.4.1 Pre-test

In the pre-test mostly personality variables and demographic variables were acquired. These were based on a scale developed by Bateman and Crant (1993), which was used to measure personal dispositions toward proactive behavior. In this research the scale was adjusted and used to assess the proactive personality of a participant. An example of this was the statement “I excel at identifying opportunities”. The personality variables were assessed by means of rating questions, thus the respondent was able to choose to which level he or she agreed with the statements. The options for rating were of the 7-point Likert Scale, seven points is the ideal amount according to Symonds (1924).

In addition to this, statements were used based on the Mini-IPIP to assess personality. This scale makes use of the International Personality Item Pool to evaluate personalities, in

(24)

this research an adjusted version was used. An example of such a statement was the

following: “I have a vivid imagination”, where the possible answers were also based on a 7 point Likert Scale in which a participant could agree or disagree to a certain level.

Having said this, an important variable that was asked in the pre-test amongst other demographic variables was the variable of tenure, which was treated as a moderator in the last propositions. Tenure was asked as an open-ended question: “How many years have you worked at the organization you currently work for?”.

4.4.2 Diary study

In this research there was a need to know if the respondent was being voiced to and what kind of message this voice contained. To acquire this knowledge, questions were asked in which the participant could give his or her perception of a co-worker’s voice. The questions that illustrated suggestion voice were based on the literature of Morrison (2011), an example was “Did your co-worker give a work-related suggestion this morning?”. Promotive voice and Prohibitive voice of a co-worker were both assessed based on literature of Liang et al. (2012). In this research their measure for these factors was adapted to a one-item measure which had been tested on 500 respondents in a pre-study. An example of such a question was “Did your co-worker voice about ways to prevent harm to the organization this morning?”, this particular question was pointed towards prohibitive voice. Just like the prior questions, the answers were in the form of the 7-point Likert scale.

Questions about OCB appeared in the afternoon survey. In this way it could be evaluated whether a certain voice content of a co-worker mentioned in the lunch survey, could be treated as an antecedent of OCB. Information was gathered considering both the affilitative and the challenging form, again this was done by means of rating questions based on the 7-point Likert Scale. An example of this was “I often change the way I work to

(25)

improve efficiency”. Likewise, an example of a statement considering the affilitative form of OCB was “I get involved to benefit this work-group”.

4.5 Data sources

The data sources used were surveys filled out by random people who worked fulltime or who were fulltime interns within Dutch companies. These were contacted by email, telephone, social networks and forums. In this manner it was attempted to reach people as random as possible.

4.6 Advantages and Limitations

There were some methodological limitations to this study. One profound limitation of the study design was that there was no list of the entire population to draw from, which is best when using the random sampling method (Saunders et al., 2009). Also, a person could suffer from selective attention, where only a limited number of signals can be identified, stored and used to control behavior. In this study the data on voice and behavior was based on the opinions and experiences of respondents, this data could therefore be biased. Another limitation of this study was due to the method of a diary study. It was difficult to make sure that participants kept filling out three surveys a day throughout a whole week, they tended to miss one or just stop filling out surveys after a while. This was not always intentional, as some participants received emails in their spam instead of in their regular inbox and therefore were not on time. Also, people who worked full time for a company tended to be busy during work hours. In some professions, participant were on the road, in meetings and at

appointments, therefore not everyone found the opportunity to fill out the surveys on time. Unfortunately, it was hard to make the surveys a priority on their work-day, especially since they conveyed a time-constraint.

(26)

Nonetheless, the study also contained advantages. As mentioned, the diary study was a form of the experience sampling method. It let researchers investigate a range of situational incidents of a psychological phenomenon, as a snapshot did not properly indicate the setting. Also, this form allowed for greater generalizability of the results, as the concepts were validated in real-life settings due to the method not being intrusive. Another valuable aspect of experience sampling was the sensitivity it had towards differences in individuals that emerged over time and different situations. Also, the bias of having to re-call knowledge was lessened, as not much room existed between the repeated surveys (Scollon et al., 2003). With the presence of anonymity and the study not being intrusive, the respondents had the freedom to be honest in their answers, which lessened the subject bias. Also observer bias was lessened, since there was one observer and thus one interpretation. In addition,

comparisons of many respondents were possible due to standardized answers. This all enhanced reliability (Hoyt, 2002).

5. Results

In the previous section the research was discussed in terms of its design and its measurements. In the following section the results of the research will be specified.

5.1 Descriptive statistics

In order to get familiar with the used sample, this section presents several of its features and gives a description of the collected data.

5.1.1 Sample characteristics

In total, 187 respondents agreed to participating in this research which lasted five days.

Every day of every participant was treated as a separate respondent, since every day measured behavioural reactions to voice that were relevant to this research. Thus, one

(27)

participant could in fact represent five respondents. With this in mind, a sample with a total of 935 respondents was expected. However, some data was not collected and not every respondent had filled out the pre-test. When assessing complete data, variables that were usable for this particular research were specifically looked at. Thus, also incomplete surveys were used if they at least contained the relevant data. The relevant data for this research was collected by the questions about tenure, the voicing of co-workers and the OCB of

participants, when only taking these questions into account a valid sample of 272 respondents remained. With this tactic, the sample was left as large as possible in order to make results more generalizable. Besides, these were the only variables of interest in this particular research model. Moreover, our dependent variable was the OCB of an employee

(participant). By excluding participants who did not fill out every question, the remaining sample would only have contained respondents inclined to show particularly good behaviour. Likewise, the missing data was often due to respondents being kept by occupational meetings and appointments. Excluding these participants would exclude particular sorts of jobs. In both cases the randomness would have diminished and it possibly would have created a bias. Therefore, it was chosen in this research to include as much relevant data as possible, resulting in the valid N of 272 respondents.

The high amount of missing data in this research was partly due to the design,

participants tended to miss one or more of the surveys in an entire diary week. Another cause for this was the survey-question inquiring whether a participant had experienced voicing of a co-worker that day. The following questions continued in this theme about the voice of a colleague, if it did not occur these questions were not asked. Hence, this follow-up data was then lost. However, this tactic was a necessity and implemented in order to diminish the irritation factor amongst participants. In this research, avoiding irritation was very important, as respondents had to be kept motivated in order to keep filling out the surveys through the

(28)

entire week.

The sample that was eventually used contained 119 (43,8%) fulltime workers or fulltime interns that were female and 153 (56,3%) who were male. The distribution of age in our sample had a main focus between 20-30 years (38,2%) and 10,7% of the sample

belonged in the middle-age category of 30-40 years. This leaves the 40-50 and 50-60 years categories containing around the 20% (19,5% & 22,8%) of the sample. Older and younger ages do not work in general, therefore these were hardly represented in this sample. Of these respondents, most had enjoyed an education level of HBO (44,9%) or WO (36,4%), the rest were educated on a lower level of MBO or high school. Over half of these respondents (52,2%) were working for their organization for 5 years or less. A longer tenure than this was less common in our sample, but existent. The cases in which respondents were working for longer than 30 years for the same organization were experienced as extreme in comparison to the rest. Yet, these cases were seen as valuable to our research since workers with such tenures are expected to show the most clear dissimilar results in their behaviour. When considering tenure as a moderating factor, these are the cases likely to produce the richest information (Table 1).

(29)

5.1.2 Data descriptions

In order to get a better understanding of the sample and the general valuations of the respondents, it is worthwhile to show some descriptive statistics of the relevant variables (Table 2). The questions that informed about relevant variables were answered using a 7-point Likert-scale (1: totally disagree – 7: totally agree), except for the questions on tenure and whether or not a co-worker had voiced that day. From these results it was visible that when in fact a co-worker voiced to an employee (respondent), this was mostly experienced as a form of suggestion (m=5,66). This voice was least recognized by the employee as a form of

(30)

prohibitive voice (m=3,96). When looking at OCB, the results pointed out that respondents generally felt that they in fact did behave beneficially towards the organization (m= 4,1098) and that this was mostly felt to be an affiliative form of behaviour.

5.2 Reliability

In this research four newly created variables were tested on their reliability, these results are also shown in table 2. To measure their internal consistency the most common measure of scale-reliability was used, Cronbach’s alpha (Field, 209, p. 674). The first variable that was tested for its internal consistency was the newly created OCB-variable. With its 8 items, it resulted in a Chronbach’s alpha of .848. This meant that this variable had a relatively high internal consistency and that the different items in the variable actually measured the same thing. Consequently, the new variable named challenging OCB was tested on internal consistency. The Cronbach’s alpha that resulted from this had a value of .803, thus

challenging behaviour proved to have a relatively high internal consistency. Hence, all four items within this variable consistently measured the challenging part of OCB. Next, the newly created variable of affiliative OCB was tested. Here a Cronbach’s alpha of 0,840 was found, again this can be considered a relatively high internal consistency. Thus, also these four items consistently measured the affiliative OCB.

Finally, a new variable was also created to portray the general voice of a co-worker, leaving the content of the message aside. This variable named voice, therefore took into account promotive-, prohibitive- and suggestion voice. Unfortunately when tested, it resulted in a low Cronbach’s alpha of 0,523. However, the three items were all essential to the

research and therefore no item could be removed. The low internal consistency could have been expected due to the fact that all three items measure a different form within voice. Yet together these items do form the general voice and in a pre-test the alpha of this measure was

(31)

above the 0.800. Accordingly, it was chosen to continue the use of this variable throughout the research.

5.3 Correlations

In this section some correlations are shown, these are visible in table 2. The first highly significant (all ps<0.001) correlation was apparent between OCB and tenure (r=-.204). This could imply that the longer a person works for the same organization, the less likely he or she is to show extra beneficial organizational behaviour. Next, a slight but highly significant (all

ps<0.001) negative correlation between tenure and affiliative OCB (r=-.236) was found. This

is a surprising result, it could mean that a person who is employed longer at the same company tends to show less behaviour that aims towards keeping the company at a steady level.

Suggestion voice had a slight positive correlation with both prohibitive (r=.204) and promotive (r=.260) voice, these correlations were small, yet highly significant (all ps<0.01). This result is surprising, it could mean that when a colleague utters a prohibitive or promotive message, a suggestion will most likely follow. In the same row in table 2, a highly significant correlation is visible between voice and suggestion voice (r=.595). This was expected and was also the case for promotive (r=.760) and prohibitive voice (r=.779) (all ps<0.01).

Prohibitive voice positively correlated with promotive voice (r=.346) in a highly significant way (all ps<0.01). This possibly indicates that when a colleague mentions a concern or a problem, that he or she also tends to talk about changes or new ideas. This correlation is unexpected and can produce interesting thoughts. The prohibitive voice also correlated positively with OCB (r=.225), which shows that it is likely that an employee will show positive organizational behaviour when his or her colleague mentions a concern or issue. This result is again surprising and therefore it will be covered in a later section. Likewise a highly significant (all ps<0.01), slightly positive correlation (r=.203) exists

(32)

between prohibitive voice and affiliative OCB, this adds an extra surprising dimension to the former result.

When looking at the fourth row, it appears that OCB also slightly positively correlated with promotive voice in a significant manner (all ps<0.05). This indicates that an employee might show more beneficial behaviour within the organization when a colleague suggests new ideas and changes, this result was in fact expected. Likewise, the highly significant (all

ps<0.01) slightly positive correlation (r = .167) between promotive voice and challenging

OCB also did not come as a surprise. It possibly shows that when an employee is being promotively voiced to, he or she will most likely tend to for instance make constructive changes. Following this, in the fifth row two very high correlations that are also highly significant (all ps<0.001) are visible between OCB and both challenging OCB (r=.864) and affiliative OCB (r=.862). However, these results were expected, when an employee shows OCB it will be either affiliative, challenging or both. The correlation between OCB and voice (r=.224) was also anticipated. This small positive but highly significant (all ps<0.001)

correlation can imply that when an employee voices, he or she displays a form of OCB. The highly significant (all ps<0.001) correlation between the challenging OCB and the affiliative OCB came unexpectedly. This could mean that when an employee shows OCB, it is not necessarily only challenging or only affiliative. It is plausible when looking at the results, that an employee mostly displays both forms of OCB, which is a fascinating result. Voice seemed to also have a significant correlation with the challenging form of OCB (r=.199), as

mentioned before this is a result that was anticipated. Conversely, this was not the case between voice and affiliative OCB. But still, also here a highly significant (all ps<0.001) correlation of (r=.188) was found. This value is visible in the last row of the table and it stimulates curiosity.

(33)

Table 2. Correlations

(34)

5.4 Regressions

In this section linear relationships between variables of voice and OCB are specified by the use of multiple regressions. This is done in order to test the formed propositions regarding these relationships and to detect a moderation effect of tenure.

5.4.1 Relationship between (different forms of) voice and OCB

In table 3 the direct relationship between the independent variable general voice and the dependent variable OCB is explored. The resulting R-square suggested that the model accounted for 5% of the variance that existed in OCB. Such a low value is common, as the model dealt with human behaviour which is mostly experienced as being unpredictable. Next, the overall effect of the model seemed highly significant (F=14,287, p>0.000), which

indicates that this linear relationship was unlikely to be due to a fluke. Following this, the beta (B=.226) indicates that when the level of a co-worker’s voice goes up by one point in the experience of the employee, that the employee’s OCB will consequently go up by .226 points. Therefore a colleagues’ voice, never mind the content of the message, seems to have a small positive effect on OCB.

Table 3

Model Regression (IV: Voice, DV: OCB)

Rsquare Adj. Rsquare F Sig.

0.05 0.047 14.287 0.000

Predictor Beta Sig

Voice 0.224 0.000

In table 4 the effects of independent variables promotive voice, prohibitive voice and suggestion voice on dependent variable OCB are shown. These predictors were put together in one single regression-model with the motivation to control for each other in order to

(35)

reduce the confounding effects of other voice-messages on the dependent variable when wanting to analyse just one relation. Since different messages can be voiced in one conversation, it seemed sensible to let these variables control for each other. The original intention was to also incorporate general voice in this regression, however this led to multicollinearity.

The R-square indicates that the model accounts for 5,7% of the variance that existed in the dependent variable OCB. The overall effect of the model was highly significant (F=5,396, p>0.001), suggesting that this linear relationship among the variables has a 0,001 chance of being due to coincidence. Subsequently, it is visible in the table that only one of the coefficients was actually significant, namely the predictor prohibitive voice showing a

positive relation (p=0.03, B=.193). Considering our expectations, this is a remarkable result. The other two voice-predictors did not turn out significant, which suggests that when for instance a suggestion is done or a co-worker propounds change, it does not necessarily affect OCB. In the case of promotive voice this result was unexpected and this will also initiate interesting discussion.

Table 4

Model Regression

(IV:Suggestion -, Promotive -, Prohibitive Voice, & Tenure DV: OCB)

0.057 0.046 5.396 0.001

Predictors Beta Sig

Suggestion Voice 0.028 0.649

Promotive Voice 0.075 0.246

Prohibitive Voice 0.193 0.003

(36)

5.4.2 Relationship between (promotive) voice and Challenging OCB

In this regression (Table 5) two predictors were taken into account, the general voice and the promotive voice of a co-worker. Both these predictors were simultaneously used in the regression in order to control for each other. The regression explored if these variables

showed a linear association with the dependent variable; challenging OCB. The model fit was not necessarily high, only 4% of the variance that existed in the dependent variable

challenging OCB was to be explained by variances in the two dependent variables. However, the total model did show a significance, most likely meaning that the linear relations in this model were unlikely to be caused by chance (F=5,641, r=0,004). However, when looking at these variables as individual predictors, neither showed a significant influence. Thus, both promotive voice (B=.037, r=.688) and voice in general (B=.171 r=.064) of a co-worker were not proven to necessarily have an influence on an employee’s challenging OCB. This means that when a co-worker speaks up in general or tries to propound a change, the employee will not necessarily want to alter constructs of the organization. This might cause interesting discussion.

Table 5

(IV:Voice, Promotive Voice, DV: Challenging OCB)

Rsquare Adj. Rsquare F _Sig.

0.04 0.033 5.641 0.004

Voice 0.171 0.064

Promotive Voice 0.037 0.688

(37)

5.4.3 Relationship between (promotive) voice and Affiliative OCB

In this part of the results, the possibility of linear relations between the predictors general voice and promotive voice with dependent variable affiliative OCB is discussed. Again both these predictors were simultaneously used in the regression (Table 6), in order to control for each other. The model explained 4.2% of the variance in the dependent variable. Following this, it is visible in the table that the overall effect of the model was significant (F=5.858,

r=0.003), thus the chances are very small that this linear relation was incidental. Moving on

to the coefficients, it was shown that the predictive value of general voice was in fact significant for the dependent variable of affiliative OCB (B=.282, r=0.002). This result was against expectations and will most likely result in interesting discussion. The opposite was the case when looking at the predictive value of promotive voice, no significance was found here (B=1.124, r=0.117). This could indicate that when a change to existing operations is mentioned, that this will most probably not lead to affiliative OCB, as was expected.

Table 6

(IV:Voice, Promotive Voice, DV: Affiliative OCB)

0.042 0.035 5.858 0.003

Voice 0.282 0.002

Promotive Voice -0.124 0.117

(38)

5.4.4 Relationship between Tenure and OCB

In this section the interaction effects of tenure are explored. In order to come to the

appropriate regressions, the direct effects on OCB were again incorporated in the final model. These effects are shown in model 1 of both table 8 and 9, the direct effects shown in the second model of tables 8 and 9 are to be ignored. Since the variable tenure was incorporated into the final model, it was chosen at this stage to also integrate tenure as a control variable in the assessment of the direct effects on OCB. This was done to see whether or not these direct effects held up when the view was broadened and tenure was included.

Firstly the actual direct effect of tenure on OCB was tested, in table 7 it is visible that there existed a significant negative relationship. Thus, the longer a person works for the same organization, the less organizational beneficial behaviour he or she will portray. Accordingly, a high tenure seems to lead to lower OCB when leaving voice aside.

Table 7

Model Regression (IV: Tenure DV: OCB)

Rsquare Adj. Rsquare F _Sig.

0.042 0.038 11.78 0.001

Tenure -0.204 0.003

5.4.5 Direct effects of (different forms of) voice on OCB with added control variable tenure The direct effect of voice hardly changed due to the added control variable of tenure (Table

8, Model 1). The beta only increased with 0.002 points whilst the significance stayed the same (B=.226, r=0.000), thus the prior model on this effect holds.

When the control variable tenure was incorporated in the regression that analysed the direct effects between promotive- prohibitive-, suggestion voice and OCB, the prior model also held up here (Table 9, Model 1). Again both prohibitive voice and tenure showed a

(39)

substantial significant linear relationship, tenure was negative and prohibitive voice was positive.

Thus, adding tenure as a control variable to the models when direct effects were considered did not affect the essence of the results and the model kept standing.

5.4.6 Interactions of tenure on the relationship between voice and OCB

Due to multicollinearity, the interaction effect of general voice could not be analysed in the same regression as the other forms of voice. Thus, in this part the moderating effect of tenure on the relationship between just voice and OCB is explored with the use of an interaction effect of tenure together with voice.

In table 8 (Model 2) the interaction effect of tenure and voice on OCB is shown, it was not significant (B=-.034, r=.555). This result was unexpected, as it most likely means that tenure does not have a moderating effect on the relationship between voice and OCB.

Table 8

Model 1: Direct effects; Model 2: Interaction (Moderation of tenure on relationship Voice & OCB)

R model 1 R model 2

0.305 0.307

Model 1 Direct effects Beta Sig

Tenure -0.207 0.000

Voice 0.226 0.000

Model 2 Interaction variables Beta Sig

Tenure -0.209 0.000

Voice 0.228 0.000

Interaction Tenure & voice -0.034 0.555

(40)

5.4.7 Interactions of tenure on the relationship between promotive -, prohibitive-, suggestion voice and OCB

In table 9 (Model 2) all three interaction variables of tenure with promotive -, prohibitive -, and suggestion voice are visible. Unfortunately, also here none of the interaction variables were significant. Consequently, it could indicate that tenure does not have a moderating influence on any of the relationships between OCB and promotive -, prohibitive - or suggestion voice.

Table 9

Model 1: Direct effects; Model 2: Interaction (Moderation of tenure on relationship Different forms of Voice & OCB)

R model 1 R model 2

0.317 0.324

Model 1 Direct effects Beta Sig

Suggestion 0.019 0.755

Promotive 0.086 0.177

Prohibitive 0.192 0.002

Tenure -0.208 0.000

Model 2 Interaction variables Beta Sig

Suggestion 0.030 0.636

Promotive 0.093 0.154

Prohibitive 0.181 0.005

Tenure -0.208 0.002

Interaction Suggestion & tenure -0.061 0.380

Interaction Promotive & tenure -0.016 0.837

Interaction Prohibitive & tenure 0.027 0.671

(41)

6. Discussion

6.1 Influence of a co-worker’s general voice on (Challenging) OCB

Here the influence of a co-worker’s general voice on an employee’s OCB is discussed. Based on previous literature, it was proposed that employees would consequently show better behavior after being voiced to in any way. This has been confirmed in the results of our research, by a substantial positive linear relation between these two variables. This supports the argument found in prior literature that due to human nature, employees will attempt to copy proactive behavior of co-workers and thus enhance their OCB (Rendell et al., 2010). Also, one can now presume that co-workers might partially regulate each other’s behaviors (Feldman, 1984).

Following this, it was predicted that a co-worker’s voice would more likely influence the challenging OCB than that it would influence the affiliative OCB. However, the results contradict this expectation. The challenging OCB did not enhance as expected as a

consequence of voice, this linear relationship was not of substantial meaning. Added to this, the actual standardized effect of voice on challenging OCB was smaller than the effect of voice on affiliative OCB. What is more, the effect of affiliative OCB was of substantial value. So this relationship could not be due to chance, unlike the relationship between voice and challenging OCB.

This is surprising, since the results had suggested that a person does in fact tend to copy his or her peers. A plausible explanation for this is that copying as a human way of learning, has its limits. An employee might indeed feel the need to copy, but being a person that does not speak up in advance, it might not lie in this person’s nature to actually show pro-active behavior. Respectively, such a person can be seen as a follower. Looking at literature concerning characteristics of leaders and followers, it appears there is a difference between a person who challenges the process and a person who embraces this change and

(42)

does the job (Baker, Mathis, & Stites-Doe, 2011). It might have been too presumptuous to anticipate a person to leap into another character and go from being a follower to being proactive. The utilized theory is most probably not incorrect, but limited. When copying a co-worker, an employee will firstly tend to stay within his or her own comfort zone and thus most likely enhance in the affiliative form of OCB.

Taking this into consideration, it might be more effective for an employee to learn from its peers in a different manner. An employee incorporating his or her own thoughts and solutions when being confronted with a problem, suggestion or idea, might be more

successful than simply copying the behavior of a co-worker. According to Butts (1982), a person learns how to deal with a non-routine problem by using previous thoughts on examples, reactions and solutions. People then tend to make the idea or problem their own and the fitting behavior might subsequently come more natural to them. The cognitive part of the learning process is then dealt with and the experiential part is left. One can then make use of these previous thoughts on solutions and reactions, and learn further by doing (Zollo & Winter, 2002). This manner of learning might make an employee more inclined to leave his or her comfort-zone and thus actually display proactive beneficial behavior.

6.2 Differing effects on OCB of diverse dimensions of voice

Based on previous literature, it was expected that an employee would behave more beneficial if a co-worker mentioned a new idea or a constructive change. In other words, the promotive voice of a co-worker was expected to enhance the OCB of an employee. This expectation is not supported by the results found in the regression, the effect that promotive voice had on OCB was not of substantial value. The theory on which our expectation was based was the idea that there is a general resistance to change but that participation is the primary method to overcome this (Dent & Goldberg, 1999). In our proposition it was expected that promotive voice would make employees participate in the idea of change and that this would rule out

The indirect effect of voicing workers on organizational success : assessing the influence of an employee’s voice on the beneficial behavior of co-workers