THE VALUE OF SOCIAL FACTORS IN
EXPLAINING KNOWLEDGE SHARING:
THE INFLUENCE OF MOTIVATION, TRUST AND COMPETITION
FACTORS ON KNOWLEDGE SHARING INTENSITY
Master thesis, MscBA, specialisation Change Management
University of Groningen, Faculty of Economics and Business
April 1, 2010
MELANIE M. DE WAAL
Student number: 1468464
Schuitemakersstraat 2-13,
9711 HW Groningen, The Netherlands
tel: +31(0)645928009
E-mail: m.de.waal@student.rug.nl
Supervisor/ university: C. Reezigt/ J.F.J. Vos
Supervisor/ field of study: (...)
2
THE VALUE OF SOCIAL FACTORS IN
EXPLAINING KNOWLEDGE SHARING:
THE INFLUENCE OF MOTIVATION, TRUST AND COMPETITION
FACTORS ON KNOWLEDGE SHARING INTENSITY
ABSTRACT
The purpose of this research is to identify the most relevant factors for explaining knowledge sharing. For this aim, this study focuses on the influence of social factors on knowledge sharing intensity. Hypotheses concerning the factors motivation, trust and competition were tested by conducting a questionnaire and interviews with change managers of a Dutch grid operator. These results were combined by triangulation. Findings of this study indicate that the factors motivation, trust and competition are important in relation to knowledge sharing. Specifically, the motivation factor extrinsic rewards and trust factor social interaction are relevant in explaining the intensity of knowledge sharing. Consistent rewards for knowledge sharing have a positive effect on knowledge sharing. This study removed some of the ambiguity, which surrounds the relationship between trust and knowledge sharing. Since, more evidence was found for an indirect relationship, than a direct relationship, between these factors. Trust influences knowledge sharing indirectly, because trust was only positively related to motivation which then influences knowledge sharing. This finding was strengthened by the positive influence of the trust factor social interaction on motivation. An unexpected finding was that the indirect negative influence of internal competition on knowledge sharing was defined by two paths. Next to motivation, competition was also found to negatively influence trust. Future research should be conducted to determine the relationships and the respective influence of the factors trust and competition on knowledge sharing. This future research should also address the limitations of this study, mainly due to reliability and validity issues.
Key words: knowledge sharing, knowledge management, motivation, trust, competition.
Acknowledgements:
3
TABLE OF CONTENTS
1. INTRODUCTION ... 4
1.1 Research Context: organisation X ... 5
1.2 Management Question and Problem Statement ... 6
1.3 Objective Statement and Research Question ... 7
2. THEORETICAL FRAMEWORK ... 8 2.1 Motivation ... 9 2.2 Trust ... 11 2.3 Competition ... 14 2.4 Conceptual Models ... 15 3. METHODS ... 17 3.1 Data Collection ... 17 3.2 Measures ... 17 3.3 Data Analysis ... 19 4. RESULTS ... 21 4.1 Descriptive Statistics ... 21
4.2 Results of Correlation Analysis ... 22
4.3 Results of Regression Analysis ... 24
4.4 Conclusion ... 29
5. INTERVIEW REPORT ... 30
5.1 Introduction ... 30
5.2 Interview Method ... 31
5.3 Interview Outcomes ... 31
5.4 Conclusion from the Interviews ... 34
6. DISCUSSION ... 35
6.1 Triangulation Results for the Hypotheses ... 35
6.2 Theoretical Explanation for the Relationship between Competition and Trust ... 39
6.3 Limitations and Future Research ... 40
6.4 Implications for Science and Practice ... 40
7. OVERALL CONCLUSION ... 42
REFERENCE LIST ... 43
APPENDIX A: RESEARCH SCOPE ... 47
APPENDIX B: QUESTIONNAIRE ITEMS... 48
APPENDIX C: FACTOR ANALYSIS ... 50
4
1. INTRODUCTION
The importance of knowledge and its management is demonstrated by social research of the last decades (Alavi and Leidner, 2001; Davenport and Prusak, 1998). This ensured that knowledge management did not become another management fad. Instead, it offers an opportunity to improve organisational performance to keep up with competition. Hence, Nahapiet and Goshal (1998) use the term intellectual capital, to emphasize the value of knowledge and the power of sharing it in relation to creating an organisational advantage.
Davenport and Prusak (1998) provide a working definition of knowledge for this research.
‘Knowledge is a fluid mix of framed experience, values, contextual information, and expert insight that provides a framework for evaluating and incorporating new experiences and information. It originates and is applied in the minds of knowers’ (Davenport and Prusak, 1998:5).
This definition makes clear that knowledge encompasses so much more than data and information. The beauty of knowledge is that it also incorporates experience and expertise, which are essential for learning. Polanyi (1966) distinguishes two types of knowledge, namely the tacit and explicit kind. "Explicit" or codified knowledge refers to knowledge that is easily transferable in formal and systematic language (Polanyi, 1966:4). Tacit knowledge is harder to capture in documents and resides in people‟s heads. Thus, knowledge is possessed and shared by individuals.
The big benefit of tacit knowledge is that it is hard to imitate by competitors (Teece, 1998). The notion that knowledge is such a precious corporate asset, also asks for careful knowledge management (Davenport and Prusak, 1998). Von Krogh (1998) states that knowledge management, through identifying and transferring organisational knowledge, is especially useful for organisations to deal with competition. Knowledge sharing is essential in increasing efficient knowledge use and its potential value (Alavi 2000).
The urgency for knowledge sharing within organisations is also recognized in light of today‟s global, changing and competitive business environment, to achieve a sustainable competitive advantage (Davenport and Prusak, 1998). Given these circumstance, many organisations need to adopt a customer-focus. Tacit knowledge sharing is positively related to a display of customer-orientation (Reychav and Weisberg, 2008). Further, organisational learning is seen as a powerful way to compete in this complex environment. For this to take place it is crucial that organisations attract and use huge quantities of knowledge to make the necessary organisational changes (Chawla and Renesch, 1995). Cummings and Worley (2001) name five essential characteristics of learning organisations in this regard, and two of them are related to knowledge sharing. Namely, information systems and Human Resource practices, by facilitating and rewarding this behaviour.
5
1.1 Research Context: organisation X1
Organisation X was formed on July 1st 2008 after the organisational unbundling of organisation Y into a network company (organisation X) and a production and supply company (organisation Y). Organisation X is the new name for the Dutch grid operator2, which is responsible for managing the gas and electricity grids. The grid consists of the network of cables and wiring, which facilitates the transport of electricity and gas to homes and organisations, and for communal lighting. In compliance with the Dutch Independent Network Operations Act (WON) 3of 2007 the legal unbundling took place July 1st of 2009. According to the WON, from the 1st of January 2011, a grid operator can no longer be part of a holding that is also the producer and supplier of electricity. The WON is an addition to the electricity and gas act of 1998, which was aimed at stimulating competition on the energy market. Regional division of energy and gas suppliers inhibited switching between providers. This act of 1998 made switching possible, which increased freedom of choice and price for customers.
Organisation X remains a company with a rich history, with almost one hundred years of presence in the energy sector. The foundation for organisation X is laid in 1915, and this first organisation was expanded through mergers and acquisitions in the energy sector to eventually form organisation Y in 1999. Organisation X, in its new form, consists of three subsidiaries; X1, X2 and X3. Subsidiary X1 is the legal owner of the grid and responsible for independent management of the public electricity and gas networks, this is the largest part of the organisation. As can be seen in figure 1, the grid of subsidiary X1 covers a large area of the Netherlands. Subsidiary X2 operates in the free market, as opposed to the regulated work domain of subsidiary X1, and is mainly concerned with complex energy infrastructures and measuring services. Subsidiary X3 is active in the public space, and its customers are municipalities and provinces, for example for street lighting.
FIGURE 1 Working Area subsidiary X1
This picture was removed by the author.
1 Sources: intranet and annual reports organisation X 2
Netwerkbeheerder
6
This fresh start, as you can call it, asked for a new direction of the board of directors. They proclaimed the ambition to become the number one service provider in the eyes of the customer. This ambition, completed with a desire to be flexible and reliable for all stakeholders, raises a huge challenge for organisation X. Since, the entire organisation has to adapt to new ways of working to become more customer-focused. In this ambition social corporate responsibility is a core theme. This theme is translated into a mission, which focuses on the responsible use of the important position organisation X has in society. Therefore, the board of directors has declared it a corporate duty to use this position for improving and contributing to this society. In order to put this ambition and mission into practice, a new strategy and matching core values were formulated. The strategy to adopt a customer-orientation is aimed at putting the customer central to all processes, actions and decisions. To deliver such top-shelve service, cooperation between departments in the chain is necessary. World Class Customer Management (WCCM) is a major program set up to achieve this collaboration in the chain. The idea behind this program is that communicating about and sharing knowledge of customers in the chain will improve customer satisfaction. This new strategic focus means a large transition that organisation X has to go through. The accompanying core values underscore this customer-focus. These values are transferred to the organisation members by organized open corporate-wide days.
1.2 Management Question and Problem Statement
The organisation context learns us that organisation X is in transition to become more customer-focused. The urgency for knowledge sharing, for organisations such as X, is addressed in the introduction. For the success of these change effort, collaboration and sharing knowledge between departments are important issues. Therefore, initiatives to set up knowledge management are started on different levels in the organisation. Technological systems for knowledge sharing are corporate yellow pages, which list expertise, a corporate wiki-page and an intranet where information can be shared. More specific examples are workshops on knowledge management, change agent reflection meetings, and knowledge lunches on specific subjects.
The importance of knowledge sharing is relevant in two ways for the departments dealing with change management. Firstly, because these departments are responsible for the change processes related to this transition to a customer-focus. Thus, implementing an effective knowledge management system has become an important part of their activities. Secondly, these change managers also have to start „practicing what they preach‟. Thus, not only do the change managers regularly implement changes concerning knowledge management, they also have to start sharing knowledge with colleagues in the field of change management. The current pace at which similar change initiatives emerge, referred to „as mushrooms‟, creates a situation where none of the change managers has a clear overview of all running projects, managed by the different departments. Thus, knowledge sharing is necessary in particular for effective and consistent change management within organisation X.
7
Considering the previous statements, this does not lead to effective collaboration on change projects to realize the ambition of becoming more customer-focused. It seems unlikely that the change managers are able to accomplish this, when they are not working together effectively. This is rather ironic, since they are responsible for managing changes related to this transition, which is founded on the principles of collaboration. Therefore, the management question is to give insight into the factors that influence the degree of knowledge sharing and provide recommendations for stimulating knowledge sharing between the change managers.
Research Scope. The management question provides the scope for this research. The four
departments with OD-consultants in-house, who engage in and are specialized in change management, are A, B, C and D.
Department A has a special position, because of its temporal mandate to guide the transition of organisation X. Further, department A is not officially included in the organisation chart, but is positioned high in the organisation, directly under the CEO. Department B is located under strategy, but has a more integral approach to change management. Department C is part of HRM, and this provides the perspective for developing OD policy, focused on growth and learning. Finally, department D is responsible for change management at the operations level. Their focus is more on operational changes, mostly process optimization.
1.3 Objective Statement and Research Question
The objective of this research is to give insight to department C of organisation X, into the relevant factors that can explain the degree of tacit knowledge sharing between change managers of the departments A, B, C and D. Based on these finding I will provide possible recommendations for managing these factors, in light of stimulating knowledge sharing at organisation X.
In order to reach this objective, I will answer the following research question in this report.
Which factors are most relevant in explaining the intensity of knowledge sharing at organisation X between change managers of the departments A, B, C and D and in what way can these factors be managed to enable more knowledge sharing?
8
2. THEORETICAL FRAMEWORK
As the research outline makes clear, this research is aimed at investigating those factors that influence tacit knowledge sharing on the individual level. In this theoretical framework I will explore these factors that explain knowledge sharing. Firstly, I will discuss the relevant types of tacit knowledge, and the concepts of knowledge management and knowledge sharing.
Expertise and experience are clearly more tacit forms of knowledge and this is the kind of knowledge that is relevant in this research. Expertise can be defined as „the relative levels of knowledge in people‟ (Ackerman, Pipek and Wolf, 2003). In this sense it describes the level of knowledge, in terms of know-what, know-how or know-who, a person has in a particular field. Know-what concerns the content or the field where a person has a lot of knowledge about. Know-how contains the personal skills, behaviour or activities of a person (Soekijad, 2005). Know-who is based on the connections people have with each other in and outside the organisation they work in. Experience deals with events and actions in the past. Lessons learned in this past are stored in our minds as experience, which given the right context can be used again to deal with new situations.
The process of management of this knowledge is depicted in figure 2, by using the knowledge management cycle of Dalkir (2005). The three stages of this cycle are knowledge capture and creation, knowledge sharing and dissemination and knowledge acquisition and application (Dalkir, 2005:43). The stages are linked by the activities of assessment, contextualization and updating. In light of the management question this research is focused on knowledge sharing. This also fits the people- or connection-based function that knowledge management fulfils (Dalkir, 2005). In this view knowledge management is aimed at connecting people in order to share knowledge. Bukowitz and Williams (2000) provide another useful description of knowledge sharing, as part of the knowledge management process. The authors distinguish contribution as one of the tactical stages of knowledge management. The contribute phase can also be described as sharing knowledge with the entire organisation to develop the knowledge base further. Thus, this stage resembles the knowledge dissemination and sharing phase of Dalkir‟s model (2005).
FIGURE 2
Knowledge Management Cycle
Note. Adapted from Dalkir, 2005.
Knowledge Capture and/or Creation
9
Kalling and Styhre (2003: 57) provide one of many definition on knowledge sharing and describe it as: „the idea that knowledge can be disseminated, transferred, diffused, shared and distributed within and between organisations, communities of practice and departments‟ despite the tacit dimensions of this knowledge. The authors distinguish three types of knowledge sharing initiatives. They make a separation between knowledge retrieval, creation and exchange. This research focuses on this sharing or exchange of knowledge between individuals. The human factor in knowledge sharing in these definitions should be emphasized. Knowledge sharing is mostly a social process with social factors influencing knowledge sharing (Søndergaard, Kerr and Clegg, 2007). This research also adopts such a social perspective on knowledge sharing.
2.1 Motivation
According to Bukowitz and Williams (2000) motivation is the key to knowledge sharing. The biggest motivational barriers for individuals to share knowledge are that it is seen as time-intensive and that the value of sharing knowledge is unclear (ibid). The importance of individual motivation for knowledge sharing is underscored by many authors (Cummings and Teng, 2006; Kalling and Styhre, 2003; Søndergaard et al., 2007).
Hypothesis 1. High motivation for knowledge sharing will lead to a high intensity of knowledge sharing
Osterloh and Frey (2000) state that two types of motivation, extrinsic and intrinsic, are necessary to motivate especially the exchange of tacit knowledge.
Extrinsic motivation. Osterloh and Frey (2000) state that employee‟s extrinsic motivation stems
from indirect need satisfaction, especially through monetary rewards. This motivates indirectly, because the received reward is the source of the need satisfaction, and not the activity itself. Organisations can use this extrinsic motivation effectively, by linking rewards to organisational goals. Individuals need a clear perception of the benefit of sharing knowledge and cooperating (Bukowitz and Williams, 2000). This means that individuals are motivated to share their knowledge if they expect a sufficient reward in return. Thus, a rewarding system focused on cooperation stimulates knowledge sharing. Otherwise, this same rewarding system will induce competition. Knowledge sharing depends on whether employees are rewarded on relative or absolute performance (Ackerman et al., 2003). Rewards based on relative performance fuels competition between individuals, teams or departments. I will further discuss this competition factor later on.
10
Hypothesis 1a. Rewards for knowledge sharing will lead to a high intensity of knowledge sharing.Intrinsic motivation. Deci (1975: 23) defines intrinsic motivation as „the amount of time spent on
a task without any apparent reward, except the activity itself. People seem to engage in activities for their own sake, out of interest, or for the pleasure or satisfaction derived from it‟. In literature on intrinsic motivation, self-determination theory is a major school of thoughts. According to this theory, individuals have three needs, competence, relatedness and autonomy, which have to be satisfied to motivate the fulfilment of certain tasks, hence knowledge sharing (Ryan and Deci, 2000). The authors emphasize that the three needs can be complementary, but this depends on the degree in which the environment supports this need satisfaction. The competence need is related to successfully executing a difficult task. Autonomy is defined as experiencing choice and a feeling of being in control of one‟s own actions. The need of relatedness is aimed at forming relationships that involve mutual respect. Lin (2007) measures intrinsic motivation for knowledge sharing by two factors, knowledge self-efficacy and enjoyment in helping others. These factors resemble the intrinsic motivation needs, as I will discuss below.
The first factor of self-efficacy implies that an individual feels capable and confident to share knowledge, which will be perceived as contributing to the organisation (Lin, 2007). Yang and Farn (2009) state that self-efficacy is an internal control factor. Next, to confidence, the absence of external factors which influence this person‟s actions is important for this internal control (ibid). This resembles the competence and autonomy needs necessary for intrinsic motivation (Ryan and Deci, 2000). Further, these authors place intrinsic motivation on the far end of the self-determination continuum, which comes with a feeling of internal control.
The second factor of „enjoyment in helping others‟ by sharing knowledge is related to altruism (Lin, 2007). Altruism is defined as those behaviours aimed at helping others with relevant organisational activities (Organ, 1988). Altruism motivates intrinsically, because the act of helping is gratifying in itself (Davenport and Prusak, 1998). Individuals engage in such altruistic behaviour, because they hope to receive respect for example (Blau, 1964). This can be linked to a need for relatedness in the self-determination theory. In this theory a feeling of enjoyment is only associated with intrinsic motivation (Ryan and Deci, 2000). This enjoyment in helping others motivates intrinsically to share knowledge and increases the intensity of knowledge sharing, because people are willing to devote more time to this activity (Remedios and Boreham, 2004).
Lin‟s (2007) research demonstrates that the two intrinsic motivation factors of self-efficacy and enjoyment in helping others influence knowledge sharing. Compared to the needs of self-determination theory both factors are important for intrinsic motivation. Therefore, this research focuses on the influence of the construct of intrinsic motivation on knowledge sharing.
11
2.2 Trust
The relevance and importance of trust in light of knowledge sharing is mentioned by many authors. According to Davenport and Prusak (1998) trust is the foundation for all knowledge initiatives, without it they will fail despite all other efforts. Others state it more lightly, but agree on the necessity of interpersonal trust for successful knowledge sharing relations (Ackerman et al., 2003, Blau, 1964; Bukowitz and Williams, 2000; Staples and Webster, 2008; Willem and Buelens, 2007).
Since there are so many views on trust, a commonly used definition of trust is adopted from McAllister (1995). The author proposes two categories of trust dimensions; affective-based and cognitive-based. Affective-based trust stems from an emotional commitment between individuals. The cognitive basis for trust concerns the choice in trustee and the trustworthiness of this person, and can also be defined as reliability. This reliability factor in the knowledge sharing relationship is used in this research, because it is the most common and straightforward dimension of trust.
Next to this dimension of reliability I have added two other trust dimensions, social interaction and reciprocity, because their relevance was demonstrated by literature review. For example, Davenport and Prusak (1998) state that trust has to grow in time and this happens mostly when people get acquainted with each other. Many authors agree on the importance of social interaction ties in order for a trusting relationship to be established for knowledge sharing (Chiu, Hsu and Wang, 2006; Søndergaard et al., 2007; Willem, Buelens and Scarbrough, 2006). Reciprocity is also mentioned as an important factor for building trust in relation to knowledge sharing (Blau, 1964; Bukowitz and Williams, 2000; Chiu et al., 2006; Davenport and Prusak, 1998).
In spite of the consensus on the importance of trust for knowledge sharing, the exact influence of trust in this relationship is not clear-cut based on research. Especially, the relationship with motivation in this regard is ambiguous. There is a large body of evidence for the direct relationship between trust and knowledge sharing. Although, there are also strong grounds to assume that motivation has a significant role in this relationship. By testing these alternative explanations for the influence of trust I will try to remove some of this ambiguity. Therefore, trust hypotheses are included in two-fold and two conceptual models were composed to display these relationships. In the following paragraphs I will explore these two options.
12
stating that the higher trustworthiness a person displays results in higher levels of resource exchange, in this case knowledge, with this person. This demonstrates that trust is essential for employees to engage in knowledge sharing activities, which is translated into the hypothesis that:
Hypothesis 2. High trust will lead to a high intensity of knowledge sharing.
In addition to this direct relationship, the importance of a willingness or motivation to share knowledge in relation to trust and knowledge sharing is also demonstrated in research literature. Often, when motivation is combined with trust to explain knowledge sharing in research, the interrelation between these variables is not hypothesized or tested (e.g. Søndergaard et al., 2007 Yang and Farn, 2008). Nevertheless, the importance of motivation in this relationship is demonstrated by a review of research on trust and knowledge sharing.
Nahapiet and Ghoshal (1998) argue that if there is trust between individuals this will lead to a higher willingness to share knowledge. According to these authors, this willingness stems from trust, because it increases the confidence that people have in the receiver of knowledge, and enhances willingness for risk taking in knowledge exchange relationships. Even authors that found evidence for the direct relationship between trust and knowledge sharing, state that trust raises a willingness to cooperate (Willem et al., 2006, Willem and Buelens, 2007). This latter study also demonstrated that incentives and trust are interrelated.
A strong argument for this interrelation between trust and motivation is made by Bakker, Leenders, Gabbay, Kratzer and van Engelen (2006). These authors state that trust on its own offers a poor explanation of the level of knowledge sharing. Although trust is an important condition for knowledge sharing, its effect is not necessarily positive. In their views it is rather the absence of trust that diminishes people‟s motivation to share knowledge with co-workers, than that the level of trust directly leads to more or less knowledge sharing. This compelling case for the potential double sword effect of trust is also made by Søndergaard et al. (2007).
Given these research findings I propose a second trust hypothesis, which states that trust has an indirect influence on knowledge sharing, because of a positive relationship between trust and motivation. This means that, more trust will lead to a higher motivation to share knowledge, which in turn will result in a higher intensity of knowledge sharing between these persons (see H1). This reasoning is continued for the second set of sub hypotheses concerning reliability, social interaction and reciprocity, which are also related to motivation.
Hypothesis 2*. High trust will lead to high motivation, and thus to a high intensity of knowledge sharing.
Reliability. This reliability factor concerns the characteristics of the sharer, as well as the recipient
13
trusting the organisation in accepting and rewarding knowledge sharing. One important factor in this perception of reliability is a feeling of safety in order to engage in a knowledge sharing relationship (Siemsen, Roth, Balasubramanian and Anand, 2009). This means that sharing knowledge with co-workers should not have negative consequences for the sharer in terms of a decline in status, image or career opportunities. This stems from the fear that people will abuse shared knowledge or take credit for it falsely (McAllister, 1995). Other aspects are the credibility of the source (Riege, 2005), competence of the other party (Willem et al., 2006) and the status of the knower (Davenport and Prusak, 1998).
This implies that the perceived trustworthiness of a person is directly related to the subsequent knowledge sharing. McAllister (1995) also draws attention to the motivation that follows from this establishment of reliability, necessary to engage in knowledge sharing. This leads to the following two hypotheses concerning reliability in relation to knowledge sharing.
Hypothesis 2a. Perceived reliability of colleagues will lead to a high intensity of knowledge sharing.
Hypothesis 2a*. Perceived reliability of colleagues will lead to high motivation to share knowledge.
Social interaction. According to Davenport and Prusak (1998) face-to-face contact is necessary to
build trust. More interaction between persons will lead to a higher perceived trustworthiness of the other person (Tsai and Goshal, 1998). McAllister (1995) argues in line with this, that interaction leads to affect-based trust, which is an important component of trust in a knowledge sharing relationship.
Furthermore, interpersonal relations are an important trust factor that influences knowledge sharing (Søndergaard et al., 2007). These personal connections provide useful access to knowledge and their holders (Cross, Parker, Prusak and Borgatti, 2004). The organisation can facilitate this by providing enough opportunities for social interaction (Wu et al., 2007).
The findings of many studies indicate that a direct and positive relation exists between social interaction and knowledge sharing (Tsai and Goshal, 1998; Tsai, 2002; Chiu et al., 2006). However, Wu et al. (2007) did not find evidence for this direct positive relation, and argue that perhaps more is needed to motivate knowledge sharing. From this finding inferences can be made about the more robust role of motivation in this relationship between social interaction and knowledge sharing. Social interaction could thus also influence knowledge sharing more indirectly, by increasing motivation first and then intensifying knowledge sharing. These two possible relationships are translated into the following hypotheses.
Hypothesis 2b. Social interaction with colleagues will lead to a high intensity of knowledge sharing.
14
Reciprocity. Reciprocity refers to the mutuality of a relationship, in this case one of knowledgesharing. According to Blau (1964), a norm of reciprocity is important in building trust and essential in knowledge sharing interactions. This is confirmed in more recent studies (Chiu et al., 2006; Bukowitz and Williams, 2000). This is related to the discussion by Davenport and Prusak (1998) of the existence of knowledge markets in organisations, where employees trade their knowledge among each other. For the provider of knowledge this means providing time, energy and effort which are scarce goods for most people. Such a sacrifice will only be made when something is promised in return. These „benefits‟ can take the form of directly shared knowledge from the recipient, or a better reputation as a willing knowledge sharer. Lin (2007) mentions future requests for knowledge that will be met regarding reciprocity.
What is striking is that reciprocity is considered important regarding knowledge sharing by building trust (e.g. Chiu et al., 2006), and as an important motivation factor as well (Davenport and Prusak, 1998). This indicates that reciprocity can lead to more knowledge sharing, by building trust, but also by providing an important incentive for this behavior. The following hypotheses represent these two possible relationships between reciprocity and knowledge sharing.
Hypothesis 2c. Reciprocity in the sharing relationship will lead to a high intensity of knowledge sharing.
Hypothesis 2c. Reciprocity in the sharing relationship will lead to high motivation to share knowledge.
2.3 Competition
According to Ackerman et al. (2003) an excessive focus on competition within organisations is the most important factor that impedes motivation for knowledge sharing. This competition between individuals should be avoided, because it will lead to knowledge hoarding instead of sharing. The way in which an organisation is structured can also have an impact in this sense. Allen, James and Gamlen (2003) state that this organisational structure can raise geographical and organisational boundaries that impede knowledge exchange. This leads to the separation of employees and may hinder necessary contact to share knowledge. In the most extreme case this results in islands of knowledge (Leonard and Sensiper, 1998). Furthermore, Ackerman et al. (2003) name formalization and status hierarchies as disincentives to knowledge sharing.
Also relevant in regard to competition are organisational politics or the behaviour of politicking. Willem et al. (2006) state that politicking has a negative effect on knowledge sharing, and possibly leads to knowledge hoarding. It is a source for opportunism, since this will lead to sharing knowledge when it is only in favour of one party. Willem and Buelens (2007) argue that competition between departments could also explain the presence of power games and the negative effect they have on knowledge sharing.
15
2.4 Conceptual Models
This theoretical framework made clear which factors are relevant in explaining the degree of knowledge sharing. These factors and their relations lead to adopting two conceptual models, this in order to test the stated hypotheses and to answer the research question. Both conceptual models incorporate the hypotheses concerning motivation and competition in an identical fashion. The motivation to share knowledge, determined by the amount of intrinsic and extrinsic motivation, is directly related to knowledge sharing. The factor of competition is deemed relevant in this research and is hypothesized to have a negative influence on motivation for sharing knowledge.
For the role of trust two different possibilities are hypothesized, which are depicted in the two separate conceptual models. Figure 3 displays conceptual model 1, where trust and the individual trust factors, reliability, social interaction and reciprocity are directly related to knowledge sharing. Whereas, in conceptual model 2 (figure 4) the trust factors first positively influence the amount of motivation to share knowledge, which then will lead to a higher intensity of knowledge sharing. Thus, what can be seen is that the construct trust has an indirect effect on the knowledge sharing intensity, through a positive relationship with motivation. In this context motivation serves as a mediator between trust and knowledge sharing.
17
3. METHODS 3.1 Data Collection
Data were collected by conducting a structured questionnaire about knowledge sharing. Unit of analysis in this study is individual change managers. Anonymity was important for filling in the questionnaire, since it involves factors such as motivation and trust. Therefore, the questionnaire consisted of general statement concerning the other departments, which were answered from an individual perspective by the respondents. Answering questions about specific persons from specific departments was assumed to affect the answers, in regard to honesty and social-desirability.
The sample of the survey comprises all change managers that were qualified as such by the contact persons per department, mostly (senior) managers, which selected the persons to conduct the survey. This resulted in a sample of 32 persons for the survey, where the number of respondents per department ranged from 1 (department C), to 8 (department A), to 10 (department B) and 13 (department D).
3.2 Measures
Measures of the relevant factors influencing knowledge sharing were adopted from previous studies on this subject. All factors were measured by multi-item scales, which were based on a five-point Likert scale (ranging from 1= strongly disagree to 5= strongly agree). For the sake of clarity all five answering options were specified in the questionnaire.
Since, the respondents are employees of a Dutch organisation, the items were translated from English to Dutch. In some cases items were also adapted to the individual level and the context. To make sure this adaption did not affect the clarity of the questions a pre-test was conducted. The pre-test group consisted of five persons, which were representatives of the sample as well as outsiders with relevant experience. This combination was made to limit possible pre-test bias and to receive objective feedback on the survey. The pre-test resulted in the elimination of some overlapping questions and corrections in the translation of statements. All items and scales in the questionnaire can be found in Appendix B.
The Cronbach Alpha was calculated for all measures to test for reliability. The .70 value was used as a cut-off score, see table 1 (Nunnally, 1978). When the Alpha of the original scale is .70 or higher, then this is considered as sufficient and items that improve this reliability will not be deleted. In table 1 sufficient Cronbach Alpha scores were emboldened. Low reliability scores indicate low inter-correlations between the items and that the items do not measure the same underlying construct. This introduces the risk of type II errors, where not all significant relationships might be found (Willem et al., 2006).
Knowledge sharing intensity is defined as the frequency with which knowledge is shared among
change managers (Willem and Buelens, 2007). This factor was measured by using 4 items adopted from these authors. The reliability score of this scale was an acceptable .72. One item concerning the sufficient amount of knowledge sharing improved the alpha to .76 if deleted.
Motivation is operationalised as extrinsic and intrinsic motivation. Extrinsic motivation is
18
reliability if deleted. The concept of intrinsic motivation was operationalised as perceived internal control (Yang and Farn, 2004) and enjoyment in helping others (Lin, 2007). To measure intrinsic motivation 6 items were adopted from these authors. This scale did not meet the reliability of .70, and scored an unacceptable .53, with no opportunity of improving the reliability by deleting items. When the items of both concepts are combined to test reliability for the construct of motivation, this scores a Cronbach Alpha of .68. This indicates that the items which compose the construct of motivation are interrelated. This sufficient reliability score justifies combining the concepts of intrinsic and extrinsic motivation into the construct of motivation.
Trust was measured by the degree of perceived reliability of colleagues, the amount of social
interaction and the reciprocity of the sharing relationship. This construct was measured by averaging the scores of the three dimensions (Willem and Buelens, 2007). The reliability factor was measured by using a 6-item scale adopted from Willem et al. (2006). One item (concerning „knowledge abuse‟) was reverse coded, in order to make ranking of responses consistent (De Vaus, 2002). This proved not to be a reliable scale, because the Cronbach alpha was just .57 and no item could improve the reliability to an acceptable level by deleting it. The social interaction element was measured by means of 4 items that concerned opportunities for personal contact facilitated by the organisation, adopted from Wu et al. (2007). This was the only reliable trust factor scale with a score of .75. This could be improved to .83 by deleting the item concerning „possibilities to contact employees during work hours‟. The final element to measure trust was the expected reciprocity results in the sharing relationship, which was measured by 4 items that were adopted from Lin (2007). This scale did not stand the reliability test, because it generated a Cronbach Alpha of .44. The combined items of the construct trust yield an initial Cronbach Alpha of .58. This alpha is improved to .64 and .68, by consecutively deleting item 32 (regarding reciprocity) and 23 (regarding reliability). This provides evidence for the interrelation between the items that compose the construct of trust. Further, this justifies combining the concepts of reliability, social interaction and reciprocity into the construct trust.
Competition refers to the degree to which competition is present between departments, which
19
TABLE 1
Reliability Scores of Variables and Constructs
3.3 Data Analysis
The data derived from the survey were statistically analysed by using SPSS. First of all, a factor analysis was performed to test the construct validity of the questionnaire, whether the items actually measure the constructs as intended. The factor analysis could not completely confirm the theoretical constructs, which means that the validity of the questionnaire was not optimal (see Appendix C). This result could have been affected by the small sample size of this study, because larger samples, of minimally 100, give more reliable results (Timmerman, 2005). Considering the questionable reliability of these results, despite of their implications for validity, the findings in this study are based on the original variables and constructs.
Since the hypotheses in this study concern the association between variables, the assumptions for correlation and regression analysis are checked. A first issue is the measurement level of the data. Strictly speaking the use of Likert-scales in the questionnaire yields ordinal data. Nevertheless, Likert-scale data may be treated as interval data, which makes parametric tests possible (Cooper and Schindler, 2003: 254).
The assumptions for parametric tests are normality, linearity and independence of the data (Cooper and Schindler, 2003). It can be assumed that the respondents are selected independently from each other. The data of the questionnaire were explored to test for the other two assumptions. A test for normality was performed by combining a p-p plot and the Kolmogorov-Smirnov test, which confirmed each other. Therefore, table 2 displays the results of the latter test.
As table 2 demonstrates, solely the variables knowledge sharing, reliability and competition display a normal distribution. This is the case when the test has a significance that exceeds 0,05, which leads to rejecting the hypothesis proposing non-normality (Norusis, 2008). However, no relationships are proposed between these variables, thus the assumptions for parametric tests are not met. Therefore, statistical analysis to test the hypotheses will at least include the Spearman Rho correlation coefficient, a non-parametric test.
Variables and constructs Items a
1. Knowledge sharing intensity 4 .72
2. Motivation 9 .68 2a. Intrinsic motivation 4 .53
20
TABLE 2:
Results of Kolmogorov-Smirnov Test for Normality
Variables Statistic Sig.
Knowledge sharing .17 .052* Reward .24 .00 Intrinsic motivation .18 .02 Reliability .15 .14* Social interaction .18 .03 Reciprocity .21 .00 Competition .11 .20*
* Variable is normally distributed at the 0,05 significance level
For further testing a regression analysis would be the preferred method of choice, since the hypotheses propose linear relationships between the variables. Despite of the non-confirmation of the variables with the assumption of normality a regression analysis was performed. This information will be used to make educated guesses on and inferences about the hypotheses. Whenever hypothesized relationships are tested, a one-tailed analysis at the 0,05 significance level will be performed. Otherwise, a two-tailed test will be used and this will be clearly stated.
To test hypothesis 2*, which proposes an indirect relationship between trust and knowledge sharing through a positive direct relationship between trust and motivation, a mediator analysis is performed. A mediator is a third variable, which stands between the relationship of a dependent and independent variable (Baron and Kenny, 1986). The construct motivation seems to fulfil this mediating function, and stands between the variables trust and knowledge sharing. Baron and Kenny (1986) propose conditions in order to test if such a relationship exists. First, there should be a direct relationship between the dependent and independent variable. Furthermore, a direct and a positive relationship should exist between the mediator and respectively the independent and the dependent variable. The final condition states that the significant relationship between the dependent and independent variable should disappear or decrease when controlled for the mediator.
21
4. RESULTS
4.1 Descriptive Statistics
From the 32 approached change managers, 27 persons filled in the questionnaire. This provides a high response rate of 84,4 %. However, one person did not fill in the last page of the questionnaire, which produced missing values for some variables and explains why the valid sample size differs between 27 and 26 per test.
Table 3 provides the descriptive statistics and frequencies that were derived from the data. Reciprocity and intrinsic motivation have rather high means around 4,00, and high scores which start at 3,25 or 3,50. Reliability shows a similar distribution, whilst displaying a lower mean and scores which are clustered around the range of 3,00-4,17. The majority of respondents had a score around the interval of 3,50-4,00. Variables with low means are knowledge sharing, reward, social interaction and competition. For example, scores for knowledge sharing cluster around 2,00-3,00, containing more than 60 % of all responses. Reward has a similar distribution, where more than 60 % of all responses concentrates around scores of 1,50 -2,50. Table 3 also shows that the respondents filled in all possible scores for reward. The low means for social interaction and competition cannot be explained by a majority of low scores. These variables show a more dispersed distribution, see table 3. Therefore, information on the mean alone is a bit misleading, and frequencies provide a more informative image.
TABLE 3
22
4.2 Results of Correlation Analysis
Table 4 and table 5 show the correlation scores between variables in this research. Hypotheses 1a, 2a, 2a* and 2b* are supported by these results. Hypotheses 1b, 2b, 2c and 2c* could not be proved, partially because of low reliability of a number of variables or because no significant correlation was demonstrated. A number of significant relationships can be deducted from table 4, which I will discuss further. Reward correlates significantly, positively and moderately high (.36) with knowledge sharing, which is proposed by hypothesis 1a. Intrinsic motivation does not correlate significantly with knowledge sharing, thus no evidence was found for hypothesis 1b. Whereas reliability is nearly a reliable variable (see methods, table 1), it is interesting to see that it correlates significantly with knowledge sharing. Especially, since this is the only trust factor that does so directly, this lightly supports H2a. No evidence is found for direct relationships between knowledge sharing and respectively social interaction (H2b) and reciprocity (H2c). A significant relation is found between reward and social interaction (.42). This partially supports hypothesis 2b*, because social interaction is a dimension of trust and in that sense related to the construct of motivation. Since, this was not hypothesized, it is striking to find that competition correlates with a number of trust factors.
23
TABLE 4
Correlations of Variables
* Correlation is significant at the 0,05 level (1-tailed) ** Correlation is significant at the 0,01 level (1-tailed)
TABLE 5
Correlations of Constructs
* Correlation is significant at the 0.05 level (1-tailed). ** Correlation is significant at the 0.01 level (1-tailed).
2a 2b 3a 3b 3c 4 1.Knowledge sharing -.05 .36* .36* .16 -.16 -.12 2a.Intrinsic motivation - -.27 .21 .19 .72* -.42* 2b. Reward - .18 .42* -.34* -.14 3a. Reliability - .31 .04 -.36* 3b.Social interaction - -.09 -.49** 3c Reciprocity - -.19 4. Competition - 2 2* 3 3* 3a 3b 3c 4 1.Knowledge sharing .39* .39* .19 .21 .36* .16 -.16 -.12 2.Motivation - - .49** .50** .36* .45 .06 -.38* 2*. Motivation (Alternative construct) - .48** .51** .34* .48** -.03 -.37* 3.Trust - - .59** .86** .24 -.47** 3* Trust (Alternative construct) - .56** .90** .13 -.54** 3a. Reliability - .21 .02 -.36* 3b. Social interaction - -.18 -.49**
3c. Reciprocity - -.19
24
4.3 Results of Regression Analysis
Tables 6-10b display the regression coefficients for all relationships that were analysed by regression. First, I will discuss the results for the variables composing the constructs of trust, motivation and competition, followed by results for the relationships between the constructs and knowledge sharing.
Motivation. The regression model that includes motivational factors (table 6, model 1), intrinsic
motivation and reward, is not capable of predicting knowledge sharing significantly (F (2, 23) = 3,23 p = .06). This model shows that reward has a significant positive relation with knowledge sharing, where intrinsic motivation does not. Model 2 demonstrates that there is indeed a significant linear relationship between the variable reward and knowledge sharing (ß=.45, < 0,05, one-tailed). This regression model can predict knowledge sharing significantly (F (1, 25) = 6,41, p < 0,05), which is depicted by an asterisk. Reward also explains the variance in knowledge sharing significantly (R²=.20, p < 0,05). In line with the correlation results support is found for accepting hypothesis 1a, that rewards will lead to a higher knowledge sharing intensity.
Trust. In order to test hypotheses 2a, 2b and 2c the variables reliability, social interaction and
reciprocity are regressed on knowledge sharing. The results in table 7a show that the regression model with the three dimensions of trust is not capable of predicting knowledge sharing (F (3, 22) = 2,05 p = .14). It becomes clear that reliability is the only significant predictor of knowledge sharing (ß= .81, p < 0,05). When reliability is regressed on knowledge sharing (table 7a, model 2), this model is not capable of predicting knowledge sharing (F (1, 25) = 3,95 p = .06). Reliability does seem a significant predictor for knowledge sharing (ß= .73, p < 0,05, one-tailed). These results thus only partially support hypothesis 2a, concerning the relation between reliability and knowledge sharing. Nonetheless, hypotheses 2b and 2c have to be rejected based on these results. This is not surprising, since social interaction and reciprocity did not correlate significantly with knowledge sharing.
Furthermore, to test hypotheses 2a*, 2b*, and 2c* the variables reliability, social interaction and reciprocity are regressed on motivation, see results in table 7a*. These results show that the regression model with the three trust factors is not capable of predicting motivation (F (3, 22) = 1,90 p= .16). Social interaction appears to be the only significant predictor of motivation (ß= .23, p < 0,05). When social interaction is regressed solely on motivation (table 7a*, model 2), this model is capable of predicting motivation to share knowledge (F (1, 24) = 4, 39 p < 0,05). This confirms that social interaction is a significant predictor for motivation. Whereby hypothesis 2b* is supported, regarding the relation between social interaction and motivation. The evidence for hypothesis 2b* becomes stronger, because a direct relationship with the motivation factor reward exists based on correlation results (table 4) and is confirmed by regression analysis (table 7b*). Despite that this direct relationship was not hypothesized, the results implicate that there is a positive linear relationship between social interaction and rewards (B= .44, p < 0.05, two-tailed). This regression model containing social interaction is capable of predicting reward significantly (F (1, 24) = 4,793 p < 0,05), as well as explaining the variance in reward significantly (R²= .17
25
TABLE 6
Regression Analysis of Intrinsic Motivation and Reward on Knowledge Sharing
Scales Model 1 ß sig. Model 2* ß sig. Intrinsic motivation -.08 . 82 Reward .42** .01 .45** .02 R² .22 .20
ß= unstandardized regression coefficient4 R²= coefficient of determination *Regression is significant at the 0,05 level, two-tailed.
** Regression is significant at the 0,05 level, one-tailed.
TABLE 7a
Regression Analysis of Reliability, Social Interaction and Reciprocity on Knowledge sharing
Scales Model 1 ß sig. Model 2 ß sig. Reliability .81 .05 .73** .06 Social interaction .04 .83 Reciprocity -.42 .29 R² .22 .14
ß= unstandardized regression coefficient. R²= coefficient of determination *Regression is significant at the 0,05 level, two-tailed.
** Regression is significant at the 0,05 level, one-tailed.
TABLE 7a*
Regression Analysis of Reliability, Social Interaction and Reciprocity on Motivation
Scales Model 1 ß sig. Model 2* ß sig. Reliability .26 .31 Social interaction .23** .08 .42** .05 Reciprocity .13 .61 R² .21 .16
ß= unstandardized regression coefficient. R²= coefficient of determination *Regression is significant at the 0,05 level, two-tailed.
** Regression is significant at the 0,05 level, one-tailed.
TABLE 7b*
Regression Analysis of Social Interaction on Reward
Scales Model 1*
ß sig. Social interaction .44* .04
R² .17
ß= unstandardized regression coefficient. R²= coefficient of determination *Regression is significant at the 0,05 level, two-tailed.
** Regression is significant at the 0,05 level, one-tailed.
4
26
Competition. Correlation table 5 shows that there are negative correlations between competitionand a number of constructs, but no significant correlation with knowledge sharing directly is found. Negative relationships seem to exist between competition and respectively trust and motivation. Therefore, these two relationships with competition are explored further by means of regression analysis, see results in table 8a and 8b. Hypothesis 3 proposes that competition has an indirect negative influence on knowledge sharing through motivation. The results of the regression analysis of competition on motivation do not confirm the support for H3 provided by the correlation results. This regression model is not capable of predicting motivation significantly (F (1, 25)= 3,0 1 p =.10). Furthermore, table 8a does not prove the existence of a direct linear negative relationship between competition and motivation (ß= -.24, p = .10). Therefore, based solely on the regression results hypothesis H3 should be rejected. When regression and correlation results are combined a more nuanced picture emerges. The correlations results state that there is a relationship, but regression analysis cannot demonstrate that competition influences motivation of knowledge sharing. Thus, the results partially prove hypothesis H3.
The striking result of negative significant correlations between trust and competition is explored further by regression analysis, see results in table 8b. This relationship was not hypothesized and therefore a two-tailed test is performed. The regression model is capable of predicting trust significantly (F (1,24) = 8,25 p < 0,05) and competition can explain the variance in trust (R²= .26). Table 8b shows that based on the regression analysis competition has a significant negative linear relationship with trust (ß= -.24, p< .05).
Even more interesting are the correlations of trust factors reliability and social interaction with competition (table 4). These relationships are explored in a regression analysis, see results tables 8c and 8d. These tables demonstrate that competition is only a significant predictor of reliability (ß = -.24 p < 0,05, two-tailed test) and not for social interaction (ß = -.43, p = .07). The model that regresses competition on social interaction is not capable of predicting this trust factor (F (1,24) = 3,69, p = .07). This negative direct influence of competition on reliability does confirm previous statements about a link between competition and trust.
TABLE 8a
Regression Analysis of Competition on Motivation
Scales Model 1
ß sig. Competition -.24 .10
R² .11
ß= unstandardized regression coefficient. R²= coefficient of determination *Regression is significant at the 0,05 level, two-tailed.
27
TABLE 8b
Regression Analysis of Competition on Trust
Scales Model 1*
ß sig. Competition -.24* .01
R² .26
ß= unstandardized regression coefficient. R²= coefficient of determination *Regression is significant at the 0,05 level, two-tailed.
** Regression is significant at the 0,05 level, one-tailed.
TABLE 8c
Regression Analysis Competition on Reliability
Scales Model 1*
ß sig. Competition -.24* .03
R² .17
ß= unstandardized regression coefficient. R²= coefficient of determination *Regression is significant at the 0,05 level, two-tailed.
** Regression is significant at the 0,05 level, one-tailed.
TABLE 8d
Regression Analysis Competition on Social Interaction
Scales Model 1
ß sig. Competition -.43 .07
R² .13
ß= unstandardized regression coefficient. R²= coefficient of determination *Regression is significant at the 0,05 level, two-tailed.
** Regression is significant at the 0,05 level, one-tailed.
Constructs. Table 9 shows the results of the regression model that includes the three hypothesized
constructs, motivation, trust and competition to explain knowledge sharing. The predictive power of this model is relatively low (R²= .17) and not significant, thus not capable of predicting knowledge sharing (F (3, 22)= 1,48 p =.28). When Betas are compared it can be concluded that motivation is the most important factor to explain knowledge sharing. Table 9 shows that only motivation has a significant direct relation with knowledge sharing (β= .61, p < 0,05), this confirms the correlation results (table 5). Table 10a provides more support for hypothesis 1 that motivation does have a linear relationship with knowledge sharing, (β= .64, p < .05). This means that motivation can predict knowledge sharing significantly. Thus, hypothesis 1 can be accepted.
28
TABLE 9
Regression Analysis of Motivation, Trust and Competition on Knowledge Sharing
Scales Model 1 ß sig. Motivation .61** .09 Trust .12 .83 Competition .01 .96 R² .17
ß= unstandardized regression coefficient. R²= coefficient of determination *Regression is significant at the 0,05 level, two-tailed.
** Regression is significant at the 0,05 level, one-tailed.
Mediator. The results of table 10a demonstrate that motivation does not serve the function of
mediator between trust and knowledge sharing. Despite support for hypothesis 2*, which proposes that trust has an indirect relation with knowledge sharing, it cannot be verified that this is due to motivation as mediator. According to Baron and Kenny, (1986) the first condition of a mediator relationship is not satisfied. Since, table 10a (model 1) shows that there is no direct linear relationship between trust and knowledge sharing. The same table shows that motivation does have a linear relationship with knowledge sharing, (β= .64, p < .05).
A moderator relationship of trust is also not in place, because it correlates with motivation as was demonstrated earlier (Baron and Kenny, 1986). Thus, the exact relationship between trust and knowledge sharing cannot be labelled. The results, as was noted earlier, do confirm the alternative trust hypothesis (H2*), that trust has an indirect impact on knowledge sharing by a demonstrated positive direct relationship with motivation. This can be concluded, because motivation does have a direct relation with knowledge sharing as well as with trust.
TABLE 10a
Regression Analysis of Trust on Knowledge Sharing with Motivation as Mediator
Scales Model 1* ß sig. Model 2* ß sig. Model 3 ß sig. Trust .52 .28 .11 .83 Motivation .64** .04 .61** .08 R² .17 .05 .17
ß= unstandardized regression coefficient. R²= coefficient of determination *Regression is significant at the 0,05 level, two-tailed.
** Regression is significant at the 0,05 level, one-tailed.
TABLE 10b
Regression Analysis of Trust on Motivation
Scales Model 1*
ß sig. Trust .67 .02
R² .20
ß= unstandardized regression coefficient. R²= coefficient of determination. *Regression is significant at the 0,05 level, two-tailed.
29
4.4 Conclusion
Although the conditions to make sound statistical decisions were not present for all variables, relationships or tests, conclusions can be made based on these results. Firstly, the direct relationship between respectively the construct motivation and the extrinsic motivation factor reward with knowledge sharing is demonstrated by these results. Thus, correlation- and regression analysis support hypothesis 1 and 1a. The relationship between intrinsic motivation and knowledge sharing (H1b) is not supported based on the results.
For the hypotheses concerning trust different results were reported. No statistical evidence was found for a direct relation between the trust factors social interaction (H2b) and reciprocity (H2c) and knowledge sharing. Partial support was found for hypothesis 2a, concerning the direct relationship between the factor reliability and knowledge sharing. Since, the regression results could not further confirm this relation found by correlation analysis. The statistical results did not further demonstrate that trust is directly related to knowledge sharing. Thus, no evidence was found to underscore hypothesis 2. The results were encouraging for the importance of social interaction in explaining motivation for knowledge sharing. Regression analysis showed that there is a direct positive linear relationship between social interaction and motivation (H2b*), as well as with rewards. Nonetheless, no evidence was found to establish an indirect relationship of the trust factors reliability and reciprocity with knowledge sharing, through a positive impact on motivation. Thus, respectively hypotheses 2a* and 2c* could not be accepted. Overall, the statistical results provide more evidence to accept hypothesis 2*, which proposes that trust has an indirect relationship with knowledge sharing, through a positive influence on motivation.
30
5. INTERVIEW REPORT
In this section I will present the interview outcomes on the subject of knowledge sharing. The interviews were used to get a grasp of relevant factors and to investigate the problem of knowledge sharing at organisation X. The basis for these interviews was the theoretical framework concerning knowledge sharing. It was assumed that especially the factors motivation, trust and competition were important, and in particular reward and intrinsic motivation. First of all, a recap of the theory underlying the content of these interviews will be provided.
5.1 Introduction
This research is focused on departments that deal with change management, therefore it can be assumed that all relevant persons have a basis of knowledge in their field. According to Davenport and Prusak (1998) knowledge consists of experience, truth, judgment and rules of thumb. In light of performing their jobs as change managers, sharing experience, visions and expertise on work projects would seem relevant. Therefore, it will be interesting to investigate how the interviewees see knowledge sharing and the level of that activity at organisation X. Regarding motivation for knowledge sharing, intrinsic and extrinsic motivation are important (Osterloh and Frey, 2000). Bukowitz and Williams (2000) argue that there are two schools of thought regarding motivation for knowledge sharing. The first school focuses on actual rewards to stimulate knowledge sharing and the other school claims that motivation has to be provided by a collective ground. At first sight both issues are at play within organisation X in influencing knowledge sharing. Intrinsic motivation is assumed to have an important role concerning knowledge sharing. An advantage of intrinsic motivation is that it makes transfer of knowledge possible if extrinsic incentives are lacking (Osterloh and Frey, 2000). However, the positive effects of intrinsic motivation can also be strongly undermined by an excessive focus on extrinsic rewards. Thus, the amount and effect of intrinsic motivation to share knowledge is largely dependent on the use of extrinsic rewards in organisations. Mutual trust is also a necessary condition for knowledge sharing, but it could also follow from the activity itself (Davenport and Prusak, 1998). Finally, Ackerman et al. (1993) state that competition has an impact on the motivation to share knowledge. The interviews are used to investigate the importance and roles of the trust and motivation factors regarding knowledge sharing according to change managers of organisation X. Assumptions will also be tested, by answering the following general questions:
