Dissertation for
MSc Advanced International Business Management & Marketing
Humanness in Indonesia:
Increasing intra-organisational knowledge sharing
January 2012
Karijn Scholte
Page | 2
Page | 3 Abstract
This dissertation describes the study on the presence of the management concept Humanness in Java and its effect on intra-organisational knowledge sharing. The management concept Humanness derived from the African philosophy Ubuntu. Both are built on the values compassion and
Page | 5 Table of contents
Chapter 1: Introduction 7
Chapter 2: Literature review on knowledge sharing and Humanness 10
2.1. Intra-organisational knowledge sharing 10
2.2. Dimensions for Knowledge sharing 12
2.3. Humanness 13
2.4. Dimensions for Humanness 15
2.5. Hypotheses 16
Chapter 3: Research methodology 21
3.1. Ontology 21 3.2. Epistemology 21 3.3. Methodology 22 3.4. Methods 23 3.5. Sources 24 3.6. Ethical issues 27
Chapter 4: Testing the measurement scales and describing the sample 29
4.1. The quality of the measurement scales: validity and reliability 29
4.2. The quality of the sample: descriptive statistics and excluding other
influencing factors 30
Chapter 5: Findings in Indonesia 38
5.1. Descriptive statistics 38
5.2. The correlation between Humanness and knowledge sharing 40
5.3. Regression analysis 44
5.3.1. Regression analyses of three levels 44
5.3.2. Overview results Indonesia 49
5.3.3. Comparison Tanzania 51
Chapter 6: Broader context and limitations 53
6.1. Interpretation of findings 53
6.1.1. Presence of Humanness values 53
6.1.2. Influence of Humanness on knowledge sharing 55
6.1.3. Possible reasons for unexpected findings 56
6.2. Limitations of study 57
Chapter 7: Conclusions and suggestions for further research 58
References 60
Page | 6
Chapter 1: Introduction
Intra-organisational knowledge sharing can enact as a good source of competitiveness in the increasingly global and competitive market (Lin et al., 2009). This dissertation describes a study on Humanness and its relation to intra-organisational knowledge sharing. Humanness is a management concept that is derived from the African philosophy Ubuntu and is based on human values. The study was conducted in Indonesia, which makes it the first empirical study on this issue outside the African context. This chapter will give a short introduction in the topic and sets out the content for the rest of this dissertation.
The current globalizing business environment forces organisations to keep developing their competencies, skills and knowledge in order to keep up with the rapidly changing demands
(Almahamid et al., 2010). Porter (1985) stresses that companies should accomplish this development faster than their competitors in order to remain competitive. According the knowledge-based view, knowledge is an organisation’s core competitive advantage (Grant, 1996; Teece, 2000). Numerous authors agree that knowledge within an organisation can be a competitive advantage which enables companies to be profitable (Pretorius and Steyn, 2005; Jiacheng et al., 2010) and respond to the changing business environment (Almahamid, 2010). Sharing knowledge within the organisation is crucial remain competitive (Chow and Chan, 2008; Jiacheng et al., 2010; Lee and Ahn, 2007) and create economic value (Lin et al., 2009).
A new field of research has emerged around the concept Humanness which has the potential to provide new insights into ways of increasing this intra-organisational knowledge sharing (Scholtens, 2011). Humanness (called Ubuntu in the African context) is a management concept which is based on the African philosophy Ubuntu. The Ubuntu philosophy is based on human values such as caring for one another, respect, dignity, tolerance and the importance of community (Mbigi, 1997). Khoza (1994) and Mbigi (1997) connected this philosophy to African management practices. These management practices are characterized by for example the importance of teamwork, the
Page | 7 This study was one of the first to examine the presence of Humanness outside the African context. Therefore, theoretical value of this research is the fact that it provides empirical meaning to theories which were up to now only based on assumptions. Providing support for the theory that Humanness is not strictly African creates new field of research. Now that is found that Humanness can be found in Java, and therefore outside the African context, it will be important to create deeper
understanding of the consequences of its presence. Humanness may be present in unexpected countries and organisations, and its presence may have implications we did not know of and thereby unnoticed affect the way organisations operate and perform.
Also this was only the second empirical study which investigates the relationship between
Humanness and intra-organisational knowledge sharing (Scholtens, 2011). The study was inspired by Scholtens’ (2011) study in Tanzania on the same subject (for a description of his research see
Chapter 2). By performing this study in a similar way, the results of both pieces of research could be compared. This comparison showed that the effect of Humanness on intra-organisational knowledge sharing in both countries is very similar thereby supporting the statement that the concept
Humanness found in Java was the similar as the one found in Tanzania.
Therefore the value of studying the Humanness-knowledge sharing relationship lays not only in providing new, empirically supported, insights but also it ensures that the management concept measured in Indonesia is indeed similar to that measured in Tanzania.
Knowing that Humanness as a management concept can indeed be found outside the African context and a positive relationship with intra-organisational knowledge sharing can be
demonstrated this might have major implications for management practices around the world. It can provide managers with new insight and understanding in how their organisation works and how it could improve its performance. It could then offer managers a novel way of managing their
organisations and stimulating knowledge sharing, ultimately offering them new ways to compete on the global market.
Creating this empirical meaning was achieved through the conduction of a survey research in
Page | 8
Research questions
The aim of this dissertation was twofold and therefore led to two research questions. The first goal was to determine whether Humanness as a management concept can indeed be found outside Africa. This was investigated by measuring the presence (or absence) of Humanness values in Indonesia and led to the following research question:
1. To which degree is Humanness as a management concept present in Indonesia? The second goal was to investigate whether the presence of Humanness is related to the attitude towards intra-organisational knowledge sharing. This relationship was studied by investigating the underlying dimensions of both Humanness and knowledge sharing and their reciprocal influence. This led to the second research question:
2. Does the presence of Humanness affect managers’ attitude towards intra-organisational knowledge sharing?
The study was conducted in Indonesia because of some cultural, historical and economic similarities with African countries. In both geographical areas culture is based on human values and
togetherness. Also both have a history with cultural influences from outside its own geographical area and have an emerging economic market. During the fieldwork of this study, 138 questionnaires were filled in by Indonesian managers. The organisations in which these managers work are
divergent in size and activities in order to represent a broader population (for details please refer to Chapter 3).
Page | 9 Chapter 2: Literature review on knowledge sharing and Humanness
Intra-organisational knowledge sharing has the ability to improve a firm’s performance and competitiveness (Almahamid et al., 2010). Although employees cannot be forced to share their knowledge (Gibbert and Krause, 2002), knowledge sharing can be motived by providing the facilitating factors that can be divided into contextual and individual (Almahamid et al., 2010). The most important facilitators are: [1] employee motivation, [2] leadership [3] corporate culture, and [4] Information technology (Lin et al., 2009).
Humanness is a management concept which is based on the African philosophy Ubuntu (Sigger et al., 2010). Ubuntu is a philosophy that originates from Africa. Both concepts are based on the values compassion and communality (Mbigi, 1997; Mangaliso, 2001). A recent empirical study in Tanzania has found that Humanness has the ability to increase intra-organisational knowledge sharing (Scholtens, 2011). This chapter will discuss the literature that is written on these topics.
2.1. Intra-organisational knowledge sharing
According to Almahamid et al. (2010) intra-organisational knowledge sharing helps an organisation to reduce costs, respond to the needs of consumers, improve business processes and increase profit. Sharing knowledge enables the firm to effectively create and enhance its products, services and processes (Almahamid et al., 2010). Lin et al. (2009, p. 25) state that knowledge sharing is ‘a form of organisational innovation that has the potential to generate new ideas and develop new business opportunities through socialization and learning processes’ of employees. They define knowledge sharing as ‘a social interaction culture, involving the exchange of knowledge, experience, and skills through the whole department or organisation’. Knowledge sharing includes both the act and actors of sending and receiving (Ardichvili et al., 2003; Van der Hooff and Van Weenen, 2004).
Communication of knowledge can occur directly, such as face-to-face conversation or workgroups, and indirectly, such as through a virtual environment. Furthermore, knowledge sharing can take place on different levels, such as individual, departmental or organisational (Lin et al., 2009). Also, knowledge can take many forms, such as tacit and explicit (Nonoka, 1994). This research includes all levels of knowledge and knowledge sharing within an organisation.
According to Gibbert and Krause (2002) employees cannot be forced to share their knowledge, but can only be motivated and be provided with the right tools. Organisations should therefore
Page | 10
resources to enable them to do so (Lin et al., 2009; Bock et al. 2005). According to Isaacs (1999) knowledge sharing can happen through dialogue. He describes a dialogue as thinking and reflecting together and at the same time having things happening within the persons themselves. He explains that dialogue is not only about talking, but it is about taking action (Isaacs, 1999). Like others (Lin et al., 2009; Bock et al. 2005) he stresses that dialogue cannot be forced, but that the conditions that enable dialogue can be created.
Apart from the actions managements take knowingly to facilitate knowledge sharing, the way a company is managed can also be a facilitator for knowledge sharing, and one management itself may be less aware of. This can be reflected in, for example, leadership roles.
Much research has been done into factors that influence knowledge sharing and the literature provides us with some determinants that the majority of researchers agree upon. Overall, the motivators influencing knowledge sharing can be divided into contextual and individual factors (Almahamid et al., 2010) (see Table 2.1 for a summary). Contextual factors include organisational culture, structure and interaction (Lin et al., 2009), leadership roles, a collaborative culture (Yang and Chen, 2007) and practical aspects such as time, resources and Information Technology (Lin et al., 2009). Individual factors derive from motivation and personal value and norms. Lin et al. (2009) state that an individual’s motivation can be both intrinsic, coming from the individual itself, and extrinsic, coming from the environment of the individual. According to Szulanski (1996) motivations to share knowledge originate from the institutional structure and the employees’ individual belief structure, which include values and norms.
Table 2.1: facilitating factors for knowledge sharing
Contextual Individual
Organisational culture, structure and interaction(Lin et al., 2009)
Motivation (Almahamid et al., 2010)
Leadership roles (Yang and Chen, 2007) Personal values and norms (Almahamid et al., 2010)
Collaborative culture (Yang and Chen, 2007) Time(Lin et al., 2009)
Resources(Lin et al., 2009)
Information Technology(Lin et al., 2009)
Page | 11 four dimensions that facilitate knowledge sharing (appendix 1). Based on his results, Scholtens (2011) has combined leadership and corporate culture into one dimension and used this dimensions
together with employee motivation and Information Technology to measure knowledge sharing. Therefore these three dimensions were used to measure knowledge sharing in this study as well: [1] employee motivation, [2] leadership & corporate culture, and [3] Information technology.
The two main reasons for using these dimension are: [1] the dimensions are already tested and have proven to be useful to measure knowledge sharing (in combination with Humanness), and [2] using the same dimensions as Scholtens (2011) enables the researcher to compare the results found in Indonesia with those found in Tanzania. Since the amount or frequency of knowledge sharing is difficult to measure, this study will focus on the willingness towards intra-organisational knowledge sharing. The next section will describe the dimensions for knowledge sharing in more detail.
2.2. Dimensions for Knowledge sharing
Employee motivation
Knowledge sharing will only occur when employees are willing to share their existent knowledge (Lin et al., 2009). Therefore an employees’ attitude towards knowledge sharing is very important.
This concept includes: Reciprocal benefits, knowledge self-efficacy, enjoyment in helping others and reputation (Lin et al., 2009).
Employees who believe in their ability to share useful knowledge with their colleagues are more motivated to do so (Lin et al., 2009). Although Lin et al. (2009) focussed on the motivation of
employees, employees and managers are very much alike when it comes to what they wish to obtain from their work and that the basic motivation forces are the same (McConnell, 2005). Therefore we may assume that this dimension and its attributes are appropriate to apply to this study based on questionnaires as well.
Leadership and corporate culture
According to Lin et al.(2009) corporate culture is the most influencing dimension for intra-organisational knowledge sharing. This means that a social-oriented climate is key to stimulate knowledge sharing in organisations. Attributes of this dimension are social networks, interpersonal trust, sharing culture, organisational learning capability and reward systems for encouraging knowledge sharing (Lin et al., 2009).
Page | 12
Also leadership is found to have an effect on the amount of intra organisational knowledge sharing. The dimension leadership includes clear and well communicated vision and goals, strong support and encouragement from top management and an open leadership climate (Lin et al., 2009). Although Lin et al.have determined leadership and corporate culture as two separate dimensions; Scholtens (2011) has combined these dimensions into one because he found that the scores for both dimensions were almost the same and that the questions for the dimensions were intertwined. For the sake of comparison with Scholtens’ (2011) research, in this study the dimensions will be combined into one as well: leadership & corporate culture.
Information Technology
Lin et al.’s research (2009) gives reason to assume a connection between the availability of Information Technology and knowledge sharing. A possible reason for this positive relation is that knowledge networks enable rapid search, access and retrieval of information and support
communication and collaboration within an organisation (Lin et al., 2009). Aspects of Information Technology are technology infrastructure, database utilisation and knowledge networks (Lin et al., 2009).
Intra-organisational knowledge sharing has been defined as both the act and the actors of sharing all kinds of knowledge within the organisation. In this study intra-organisational knowledge sharing is measured as the willingness to share knowledge, which is influenced by contextual factors
(leadership& corporate culture and Information Technology) and an individual factor (employee motivation). The next section will give more insight in the concept Humanness and the way it can be measured.
2.3. Humanness
Page | 13 therefore the management concept could be present in other continents as well. They proposed to rename the concept into Humanness for a non-African context and have developed a measurement tool to examine the presence of the concept. From now on this study will refer to the concept with the name Humanness.
The Humanness philosophy describes values related to human engagement and interdependent relations (Sigger et al., 2010). Recently Sigger et al. (2010) defined Humanness as a management concept, which is reflected in the people and processes within an organisation giving it the potential to influence the organisation’s performance or competences. A recent empirical study in Tanzania has confirmed a direct relation between this management concept and intra-organisational knowledge sharing (Scholtens, 2011).
The concept of Humanness is complex and has been defined in different ways. For example, Mangaliso (2001, p.24) defines it as a “pervasive spirit of caring and community, harmony and hospitality, respect and responsiveness-that individuals and groups display for one another”. Nussbaum (2003, p.21) describes Humanness as “the capacity in African culture to express
compassion, reciprocity, dignity, harmony, and humanity in the interests of building and maintaining community.”
Page | 14
2.4. Dimensions for Humanness
Compassion
The value compassion describes one’s ability to understand the dilemma of others and feeling the urge to help them (Poovan et al., 2006; Sigger et al., 2010). This value is based on the belief that all human beings are connected to each other and share a communal responsibility for each other. Because people with this value believe that sharing and giving is the only way to receive help in return, they have a high willingness to help members within and outside of their community (Poovan et al., 2006). For this reason Poovan et al. believe that Compassion is the basis for a culture of caring and sharing.
Respect/dignity
These values where originally introduced as two separate values by Mbigi (1997). However,
following other academics (Sigger et al., 2010; Broodryk, 2006; Poovan et al., 2006; Scholtens, 2011) they can be combined into one dimension because the values are highly interlinked and enable readers to compare the results of this research to those of previous and possibly future research. Respect is in the first place about tolerating and valuing other people as well as their opinions and ideas. These opinions or ideas can be related to work or ethnical or religious matters. Also respect towards elderly people, authority and other people fulfilling their tasks for the wellbeing of the community is highly valued. Mutual respect in relationships creates trust and encourages people to share (Scholtens, 2011). Respect/dignity requires people to behave respectful towards those in authority and enables them to become dignified through this respect (Poovan et al., 2006)
Solidarity
Solidarity is developed through the combined effort of individuals in the service of their community (Poovan et al., 2006). This value is based on the belief that the needs of the community are more important than those of the individual. The goal is to form a community that works together to reach communal goals and in which is no room for selfishness and competitiveness. This value can be expressed collectively through singing, effort at work, initiation rites, celebrations, rituals and family life (Nussbaum, 2003).
Survival
Page | 15 Survival depends on ‘brotherly’ care rather than on individual self-reliance (Poovan et al., 2006, p.18). Consistent with Solidarity this concept assumes feelings of responsibility for others and combined efforts to reach mutual goals (Poovan et al., 2006).
As briefly mentioned, Scholtens (2011) recently implemented a study on Humanness and its relation to intra-organisational knowledge sharing, and found a positive relation between the two concepts. The implementation of the study took place in Dar es Salaam, Tanzania, were managers were asked to participate in the study. Scholtens (2011) was the first to use Sigger et al.’s (2010) measurement scale for Humanness, which is based on four dimensions: [1] compassion, [2] solidarity, [3] survival and [4] respect/dignity. In order to measure knowledge sharing he adjusted Lin et al.’s (2009) three dimensions: [1] employee motivation, [2] leadership & corporate culture, and [3] Information Technology. Scholtens (2011) found that these humanness dimensions are positively related to the knowledge sharing dimensions mentioned above. This study performed the same investigation in Java. This way the findings can be compared
Humanness is introduced as a management concept that is derived from the African philosophy Ubuntu. Both concepts focus on human values such as compassion and solidarity. Until now, Humanness was perceived as strictly prevailing in Africa. Sigger et al.(2010) however stated that it may be found in other contexts as well. The four main dimensions of Humanness are: [1] compassion, [2] solidarity, [3] survival, and [4] respect/dignity.
2.5. Hypotheses
Four dimensions for Humanness have been introduced: compassion, solidarity, survival and
respect/dignity. Also the three most important knowledge sharing facilitators have been introduced: employee motivation, leadership & corporate culture and Information Technology. Combining these constructs and dimensions leads to several expectations (i.e. the impact of Humanness on
knowledge sharing).
Page | 16
In order to structure this investigation several hypotheses are generated. The expectations are divided into three levels; each level looking more in-depth at the relation between Humanness and knowledge sharing.
Level 1: Humanness and knowledge sharing
The first level assumes a positive relationship between Humanness and knowledge sharing. This relation is the basis for the study.
Hypothesis 1: Managers that score high on Humanness values will show a more positive attitude towards intra-organisational knowledge sharing compared to those who do not score high on Humanness values.
Level 2.a: Dimensions of Humanness and knowledge sharing
The second level goes a little further in-depth and studies the relation between the different Humanness dimensions and knowledge sharing. This level indicates which Humanness dimensions are most related to knowledge sharing. The expectation is that they are all positively related to knowledge sharing.
Hypothesis 2a: Managers that score high on Compassion will show a higher score on Knowledge sharing than managers that score low on Compassion.
Page | 17 Hypothesis 2c: Managers that score high on Solidarity will show a higher score on Knowledge sharing than managers that score low on Solidarity.
Hypothesis 2d: Managers that score high on Survival will show a higher score on Knowledge sharing than managers that score low on Survival.
Level 2.b: Humanness and dimensions of knowledge sharing
Also the effect of the Humanness construct on each knowledge-sharing dimension was investigated.
Hypothesis 3a: Managers that score high on Humanness values will show a higher score on employee motivation than to those who do not score high Humanness values.
Page | 18
Level 3: Humanness dimensions and knowledge sharing dimensions
The third level is even more specific and investigates which Humanness dimensions influence each knowledge sharing dimension. This results in many relations that need to be investigated.
Solidarity Leadership & corporate culture Compassion Survival Respect/dignity Employee motivation Information Technology
Dimensions of Humanness and employee motivation
Hypothesis 4a: Managers that score high on Compassion will show a higher score on employee motivation than managers that score low on Compassion.
Hypothesis 4b: Managers that score high on Respect/dignity will show a higher score on employee motivation than managers that score low on Respect/dignity.
Hypothesis 4c: Managers that score high on Solidarity will show a higher score on employee motivation than managers that score low on Solidarity.
Hypothesis 4d: Managers that score high on Survival will show a higher score on employee motivation than managers that score low on Survival.
Dimensions of Humanness and Leadership & corporate culture
Hypothesis 5a: Managers that score high on Compassion will show a higher score on Leadership and Corporate culture than managers that score low on Compassion.
Hypothesis 5b: Managers that score high on Respect/dignity will show a higher score on Leadership and Corporate culture than managers that score low on Respect/dignity.
Hypothesis 5c: Managers that score high on Solidarity will show a higher score on Leadership and Corporate culture than managers that score low on Solidarity.
Page | 19 Dimension of Humanness and Information Technology
Hypothesis 6a: Managers that score high on Compassion will show a higher score on Information Technology than managers that score low on Compassion.
Hypothesis 6b: Managers that score high on Respect/dignity will show a higher score on Information Technology than managers that score low on Respect/dignity.
Hypothesis 6c: Managers that score high on Solidarity will show a higher score on Information Technology than managers that score low on Solidarity.
Hypothesis 6d: Managers that score high on Survival will show a higher score on Information Technology than managers that score low on Survival.
The above hypotheses all have the same aim: investigating the relation between Humanness and intra-organisational knowledge sharing. They do so in three levels, each going more in-depth, starting with the relation between the two constructs and ending with the relations of each
Page | 20
Chapter 3: Research methodology
The previous chapter has explained the aim of the study: investigating the presence of Humanness in Indonesia and its relation with intra-organisational knowledge sharing. Now it is time to clarify how the study was conducted in order to reach this goal. This will be done in five steps: [1] ontology, [2] epistemology, [3] methodology, [4] methods, and [5] sources.
The starting point of all research is defining one’s ontology position (Grix, 2002). After the
ontological position is determined, the epistemological and methodological positions logically follow. Finally, based on the ontology, epistemology and methodology, the methods and sources that were used in this study, can be described.
3.1. Ontology
Ontology is concerned with what is out there to know. It states the researcher’s view on the nature of reality (that what is being measured) (Reissner, 2011). This includes: what exists around us, what it is made of, and how these parts interact with each other (Grix, 2002). One’s ontological position answers the question: “what is the nature of the social and political reality to be investigated?” (Hay, 2002, p.63).
This study takes the perspective of the first: objectivism which assumes that social phenomena and their meanings exist independent of social actors, (Grix, 2002). Therefore the research is based on the assumption that what is being measured is existing on its own, independent of influences of social actors.
3.2. Epistemology
Page | 21 influenced by the researcher and is quantitative of nature. This perspective enables the researcher to use deductive reasoning to test the theory presented.
Where the different ontology perspectives diverge in what social reality is; the epistemology perspectives differ in how we can know about this social reality. Logically this results in different approaches for methodology (how we can acquire that knowledge), methods (which procedures we can use to acquire it) and sources (which data we can collect) (Grix, 2002). Knowing that this study is based on the objectivism and positivism perspectives, we can now proceed to the methodology.
3.3. Methodology
This study examined firstly whether Humanness as a management concept can be found in non-African companies. Then the relationship between the presence of Humanness and intra-organisational knowledge sharing was examined. In order to save time and effort, data on both concepts were collected at the same time. The following sections will describe which data was used to answer the research questions, how the data was collected and how it was analysed.
Empirical study
Although theories on Ubuntu are numerous, so far there have been few attempts to actually measure the presence of Ubuntu as a management concept. Only one empirical study (Scholtens, 2011) has been conducted on the presence of Ubuntu and its relation to knowledge sharing. So far, no research has been done on Ubuntu or Humanness outside Africa. As a result theories and expectations are developed, but tests on the actual presence of Humanness and its practical consequences are scarce. This empirical study began to fill this gap. The benefit of empirical studies is that they enable researchers to verify existing theories (Zikmund et al., 2010). In order to do so quantitative data was collected and analysed. One of the advantages of the use of a quantitative research is that the researcher is able to remain more distant in order to increase objectiveness. Also it allows a larger sample size which results in conclusions that are more generalizable to other situations (Zikmund et al., 2010).
Geographic scope
Page | 22
Africa is that both its economies are still in a developing phase and their position on the global market is yet to be acknowledged. Also initial research on the Indonesian culture gave reasons to believe that some key values might be similar to those in Africa. The island Java in particular has been chosen because it holds the most vibrant economic environment of Indonesia and
geographically offers a secluded and manageable population. Finally, a practical reason for choosing Java is the existence of an established network of contacts in the area which makes the area more accessible.
3.4. Methods
Data collection
The data for this study were obtained through conducting a sample survey using questionnaires. Advantages of surveys are that they allow ‘quick, inexpensive, efficient and accurate’ information about a larger population and can provide the researcher with a flexible method and valuable information (Zikmund et al., 2010, p. 187). The questionnaire was based on existing questionnaires for both Humanness (Sigger et al., 2010) and knowledge sharing (Lin et al., 2009), which were used in Scholtens (2011) research on Ubuntu in Tanzania. The advantage of using this questionnaire is that it has been validated and proven to be appropriate to measure knowledge sharing and Humanness. Also they give the researcher the opportunity to gain maximum insight with limited time and resources. Finally, a major advantage of using this questionnaire is that the results of this research can be compared with those of Scholtens’ (2011) research in Tanzania.
The existing questionnaire was translated into Bahasa Indonesia which is the official language in Indonesia. This was done by two separate independent translators, after which both versions were compared and finally combined into one translated version to ensure a good translation. Normally the questionnaires would be tested on a trial group, which should consist of approximately similar participants as the actual sample, but is not required to be a statistical sample (Zikmund et al., 2010). In this case however, the questionnaire was not tested because it should not be adapted to ensure maximum gain of comparability with Scholtens’ (2011) research.
Page | 23 enough; [4] completeness, which refers to the extent to which the data includes all necessary
information.
3.5. Sources
Sample and distribution
The population of this study comprises of small to large enterprises that have an establishment in Java and have the aim to make profit. This population is mainly chosen because it is similar to the population that has been used in Tanzania. NGO’s and governmental institutions were excluded from the study because they are expected to have fewer similarities with commercial organisations.
The exact size of the total population is hard to determine, because it is difficult to get information on registered companies and a vast amount smaller companies is not even registered. Since Java has 114 million inhabitants, there might be over a half million companies located in Java. Because the population size is unknown it is difficult calculate a precise required sample size (Zikmund et al., 2010). Thomas (2004) suggests a solution for this situation which is using a sample of at least 200 when the population size is unknown. Although this was the aim when going to Indonesia, this goal was not achieved due to difficulties with finding managers that were willing to participate in the allocated time. The final number of 138 questionnaires is somewhat small for a population that is possibly as large as 500,000. This may influence the generalizability and thereby the quality of the results. However, as will be discussed in Chapter 4, when looking at the participants’ characteristics (age, gender, and sector and company size) the sample does seem to be a good representation of the population.
Because of limited time and scope the questionnaires were distributed to managers of the sample companies. As a result the data gave a good reflection on the perspectives of the managers, but might not be representative of the perspectives of employees.
The 253 sample organisations were chosen based on their fit within the population and their accessibility. Within each company a participant was sought that has knowledge of the subject, is in the position of exposing this information and willing to participate (Miller et al., 2011).
Page | 24
a research like this. In this sense the sample is self-selected rather than selected statistically, which might influence the representativeness of the entire population. However given that the participants’ characteristics are representative of the population and that multiple networks were addressed, there is no reason to assume that this influenced the results significantly. Also normality tests were applied on the dataset which showed that the data for all questions show results close to a normal distribution. Therefore, the lack of randomness of the sample does not seem to harm the quality of the results (Field, 2009).
At first the questionnaires were distributed through email. The advantage of sending the
questionnaires by email is that it saves time and effort of the researcher, and is more anonymous for the participant (Cooper and Schindler, 2003). When participants remain anonymous while filling in the questionnaire they might feel more secure to answer truthfully (Cooper and Schindler, 2003,). However, when the response rate remained low the distribution manner was changed into personal delivery. The advantage of distributing questionnaires in person is that the respond rate is likely to be higher (Cooper and Schindler, 2003). A disadvantage of distributing the questionnaires in person is that the participants are less anonymous; although their names were not recorded, the
participants may have felt less secure to answer freely. Also the participant’s answers might be influenced by the presence of the researcher. However, the researcher tried to diminish this effect by explaining that there are no right or wrong answers and by allowing the participant to decide where and when to fill in the questionnaire. Most questionnaires were not answered in the presence of the researcher. Even though the questionnaires were delivered in person, permission to do so was asked in all cases to make sure that the managers did not feel pressured to participate.
Other efforts to increase the response rate were providing a cover letter, selecting the participants carefully, and making the questionnaire interesting, well designed and easy to answer (Zikmund et al., 2010). Finally after a slow start, 138 complete questionnaires filled in by managers from 75 different organisations formed the dataset. Since a total of 253 questionnaires were distributed the response rate is 54.5 per cent. Overall the sample seems to be an accurate representation of the population. Chapter 5 will provide more insight in the sample characteristics (i.e. participants’ age and gender, and the company’s sector and size).
Measurement scales
Page | 25 Knowledge sharing was measured by the determinants [1] employee motivation; [2] leadership and corporate culture, and [3] Information Technology (Scholtens, 2011).
The questions in the questionnaire were answered using a five-point Likert scale (ranging from 1=strongly disagree to 5=strongly agree and 3=neutral). The values that derive from this scale were treated as interval. Each determinant was measured using several questions to improve reliability. The next section will describe how the collected data was analysed.
The dimensions were difficult to measure objectively; therefore measurement was based on the participants’ perceptions. This possibly influences the quality of the study because the managers may be biased or have the tendency to overpraise their organisations. However at this point there is no reason to believe that this has significant influence on the results.
Procedure
Before planning the analysis procedure for a research the level of data should be determined (Field, 2009). When it comes to questionnaires, and Likert scales in particular there are no clear rules whether the data should be treated as of an ordinal or interval level. Field’s (2009, p8.) for example states that data that is based on the participants’ subjective perception should be treated as ordinal. Others however argue that it may be treated as interval (e.g. Alberty and Mihalik, 1989). Labovitz (1967, p.152) states that treating Likert variables as interval allows us to use “more powerful, better developed and more clearly interpretable statistics” which offsets the shortcoming of not having precisely an interval scale. Because too much information will be lost using statistics test for ordinal data in this study the data is treated as interval.
The analysis of the data started with the assessment of the validity and reliability of the
measurement tools (scales). Validity is the accuracy of a scale to measure what it is supposed to measure and is appropriate for the study at hand (Zikmund et al., 2010). The reliability of the
measurement tools was computed with the use of Cronbach’s alpha. The Cronbach’s alpha measures ‘the degree to which the instrument items are homogeneous and reflect the same underlying
construct(s)’ (Cooper and Schindler, 2003, p. 237). Also, the means and standard deviations of both the two main concepts Humanness and knowledge sharing, and their underlying determinants were calculated to determine the value of the measurement tools.
After the previous tests have proved that the scales are valid and reliable, a correlation test could assess whether there are relationships between two variables (Field, 2009). The Pearson correlation test was applied to find out whether there are relationships between: [1] the main concept
Page | 26
knowledge sharing and the dimensions of Humanness, and finally [4] the dimensions of Humanness and the dimensions of knowledge sharing.
Next, multiple linear regression models were used to determine whether the relationships that were found previously are causal of nature. Also this test showed whether the found correlations are positive or negative of nature. Again this was done for the relationship between the two main concepts, for the relationship between each concept and the determinants of the other and the dimensions of each construct with the dimensions of the other construct. Linear regression models were used with the assumption that the relationship between the dependent and independent variable can be approximated by a straight line (Bowerman et al., 2009). Multiple regression models are used when there are several independent variables (Bowerman et al., 2009).
3.6. Ethical issues
All participants of the research project partook on a voluntary basis and gave informed consent. This means that the participants shared their willingness to participate after they were fully informed on the details of the study, including: its goal, the used method, the procedure, the advantages and the advantages. The participants were asked if they fully understood what was told and if they had any questions. The help of a translator was used to make sure that the communicated information was clear for all parties. To avoid pressure to participate by delivering the questionnaire in person, in all cases the participants’ permission to bring the questionnaire in person was asked by a third party before actually doing so. This was done either through email, telephone or a shared acquaintance (such as a family member or colleague). Also the participants were informed that they could withdraw their consent at any time. After the questionnaires were delivered they were only collected if the participants gave their permission to do so. Although the participants did not give written consent one might argue that filling in and giving back the questionnaire can be seen as giving consent.
Page | 28
Chapter 4: Testing the measurement scales and describing the sample
Now that the research methodology is set out we can continue with the findings. However, before statements can be made based on the data of this study, it should be assessed whether the used measurement scales have been adequate to measure what needed to be measured. Therefore both the validity and reliability of the measurement scales will be assessed (see Section 4.1). Furthermore the sample will be described in descriptive terms. Finally, the influence of certain control variables on the specific scores will be assessed to make sure that the data is not affected by external factors (see Section 4.2).
4.1. The quality of the measurement scales: validity and reliability
First the reliability of the measurement scales will be assessed, because measurement scales can only be valid if they are reliable. The reliability will be assessed by measuring the scales internal consistency. The scale’s internal consistency is the degree to which the questions measure the right dimension. This can be measured by calculating the Cronbach’s Alpha. The Cronbach’s alpha is represented as a number between 0 and 1 and should be as high as possible, but at least bigger than 0.6. In this research the Cronbach’s alpha will be calculated for each of the dimensions and also for each of the constructs.
Validity is a measure of the accuracy of the measurement scales. The convergent validity indicates the relationship between two measures of the same construct. When the convergent validity is high the two measures are related and therefore are appropriate conceptualizations. This test will be performed by applying a Pearson correlation test. To test the validity of the scales question needs to be answered: Are the questions measuring the same dimension?
Humanness
The Cronbach’s alphas for each of the Humanness dimensions show good results (Table 4.1). All Cronbach’s Alpha’s are above 0.7 which is more than acceptable.
Page | 29 Because the internal consistency is good and the validity is acceptable we can now combine the questions for each Humanness dimension and use the dimension scores as new variables.
Table 4.1: Cronbach’s alphas for Humanness dimensions Dimension Cronbach’s Alpha N of items
Compassion .736 8
Respect/dignity .821 10
Solidarity .718 6
Survival .799 9
Knowledge sharing
The Cronbach’s Alpha’s for the knowledge sharing dimensions are acceptable to good (Table 4.2). The internal consistency for the employee motivation dimension is not very high, but still acceptable. The internal consistency for both Information Technology and leadership & corporate culture are good.
The Pearson correlation tests for the knowledge sharing dimensions reveal significant correlations between all questions within the three dimensions of knowledge sharing (appendix 3.1-3.3). Therefore we may assume that the questions for each dimension are indeed measuring the same dimension. The questionnaire has an acceptable internal consistency and good validity. Therefore, the questions for each knowledge sharing dimension will be combined into a new variable.
Table 4.2: Cronbach’s alphas for knowledge sharing dimensions
4.2. The quality of the sample: descriptive statistics and excluding other influencing factors
It is now confirmed that the measurement scales are appropriate to use for this study. However, before continuing with the analysis of the data, the sample will be described.
Also the influence of participant’s characteristics (control variables) on the scores on the Humanness and knowledge sharing dimensions will be determined. The reason for this is to make sure that these
Dimension Cronbach’s Alpha N of items
Employee motivation .641 4
Information Technology .723 3
Page | 30
factors do not influence the level of Humanness and knowledge sharing and thereby affect the results. If this is not the case it will increase the value of the results.
These tests were also performed on Scholtens’ (2011) data from Tanzania by Sigger et al.(2010). The results for both data sets will be compared to see if there are any differences between the two countries.
For this study 138 complete questionnaires have been collected. All questionnaires were filled in by persons with a managerial function and an Indonesian nationality. For each participant a few factors were regarded: the gender and age of the participant and the size of the company and the sector in which it is active. The size of the company was measured by the number of employees.
Gender
As Table 4.3 shows the number of male participants is considerably higher than the number of female participants. We may argue that this is a proper representation of the population because in Indonesia (including Java), the majority of the managers is male. Although the number of female managers has been increasing slightly in the recent years, male managers are still the norm. It is notable that this division is very similar to that of the Tanzanian sample which has a division of 72 per cent male and 28 per cent female.
Table 4.3: Gender participants frequency Percentage
Male 97 70.3
Female 41 29.7
Total 138 100.00
Next, the influence of gender on differences in Humanness scores is assessed. This will be done by applying an independent-samples T test (appendix 5.1.1).
Page | 31 For the dimensions Compassion and Solidarity, no significant value was found for the Levene’s test. Therefore we may assume that the variances are approximately equal. The differences found in the t-test are also not significant.
This leads to the following conclusion: the scores on all four Humanness dimensions range from moderate to high and there are no significant differences between the scores of males and females. Sigger et al. (2010) have applied this test to the Tanzanian sample as well, which resulted in the same conclusion: no significant differences between the scores of males and females can be found.
Table 4.4: Gender and Humanness dimensions
gender participant N Mean Sig. (Levene’s test) Sig. (t-test)
Compassion male 97 4.1443 .273 .698 female 41 4.1128 Solidarity male 97 3.5722 .206 .349 female 41 3.4756 Survival male 97 4.2612 .030* .166 female 41 4.1626 Respect/dignity male 97 3.8392 .036* .573 female 41 3.7902 *Significant at a level of .05
Page | 32
Table 4.5: Gender and knowledge sharing dimensions
gender participant N Mean sig (Levene’s test) Sig (t-test)
Employee motivation male 97 4.1778
.589 .425
female 41 4.1098
Leadership & corporate culture
male 97 4.0902
.329 .739
female 41 4.0579
Information Technology male 97 3.2818
.662 .195
female 41 3.0813
*Significant at a level of .05
Company sector
The sector in which the participating companies are active might influence the scores on Humanness and knowledge sharing. Table 4.6 shows that most participating companies are active in the financial, consumer goods and production sectors. There might be two reasons for this. The first is that the responses are closely related to the available network present in Java, which in this case has a focus on these sectors. The second is that these sectors are some of the largest in Java (Central
Intelligence Agency, 2007).
Table 4.6: Sector
sector Frequency Per cent
non-governmental 2 1.4 tourism 3 2.2 financial 36 26.1 consultancy 6 4.3 technology 8 5.8 consumer goods 31 22.5 legal 1 .7 production 51 37.0 Total 138 100.0
Page | 33 Sigger et al.(2010) did the same for the Tanzanian sample. Also, some sector groups (e.g. tourism and legal) are too small to be assessed. The legal sector is combined with the financial sector because the sector is assumed to have more similarities with the financial sector than the other sectors.
Table 4.7 shows that the mean scores for both groups are very similar. All mean scores are high. Levene’s test for equality of variances however shows a significant score for compassion. Therefore we have to reject the null hypothesis and may not assume that the variances in scores of the different sectors are equal. Nonetheless, the independent sample t-test shows no significant differences between the compassion scores given by the two sector groups (Appendix 5.1.3). The Levene’s test is not significant for solidarity, survival and respect, which means that for these scores we may assume that the variances between the sector groups are approximately equal. However, the independent samples t-test does indicate a significant difference for the dimension respect. This means that there is a significant difference in scores on the dimension respect between financial and non-financial organisations. Nonetheless, the difference in mean scores is very small and sector groups have a high mean score for respect. Therefore we may assume that the difference is negligible and does not have a significant impact on the results. This is the same as in the
Tanzanian sample, where a significant difference of negligible size was found for the dimensions respect (Sigger et al., 2010). Likewise for the other dimensions no significant differences were found between financial and non-financial organizations (Sigger et al., 2010).
Table 4.7: Sector and Humanness dimensions
Sector organisation N Mean Sig. (Levene’s test) Sig. (t-test, 2-tailed)
Page | 34
Again the means for financial and non-financial organisations are very similar (Table 4.8). Therefore it does not come as a surprise that the Levene’s test for equality of variances shows no significant differences in variance between the two groups. The independent samples t-test however does show a significant for leadership & corporate culture (Appendix 5.1.4). The test does not indicate significant differences for the other two knowledge sharing dimensions: employee motivation and Information Technology. This leads to the following conclusion: financial and non-financial
organization do not show significant differences in scores for employee motivation and Information Technology. However whether the organization is active in the financial sector or not does influence its score on Leadership & corporate culture. However, because both means are high and do not differ very much, it may be assumed that the difference is negligible and does not have a significant impact on the results.
Table 4.8: sector and knowledge sharing dimensions sector organisation N Mean Sig. (Levene’s test) Sig. (t-test, 2-tailed) Employee Motivation financial/legal 37 405 .342 .107 other 101 4.20
Leadership & Corporate Culture financial/legal 37 3.91 .217 .020 other 101 4.14 Information Technology financial/legal 37 3.34 .137 .304 other 101 3.18 Age
The numbers of participants for each age group are more evenly spread (Table 4.9). As expected, the largest group is the slightly older one. In Java respect comes with seniority and experience.
Page | 35 Table 4.9: Age (in years) of participants
Age Frequency Percentage
<30 35 25.4
30-39 31 22.5
40-50 52 37.7
50< 20 14.5
Total 138 100
In order to test whether the age of the participant affects the scores on Humanness and knowledge sharing, a one-way between groups ANOVA test is performed. Also the homogeneity of variance test is done.
First the effect of age on the scores on Humanness will be assessed (appendix 5.2.1). When looking at the means for the separate age groups it is visible that they are very close. Therefore it does not come as a surprise that the homogeneity of variance test does not show any significant differences in variance. Hence we may assume that variances in the scores given by all age groups are
approximately equal. The ANOVA test supports this lack of significant difference. This means that we may assume that age does not influence the scores on Humanness. The Tanzanian sample (Sigger et al., 2010) also shows no significant differences in scores between the different age groups.
Next, the influence of age on the scores on knowledge sharing will be investigated (appendix 5.2.2). Again, the means for the different age groups are very similar. The homogeneity of variance test does not result in significant differences in variance. Likewise, the ANOVA test does not result in any significant differences between the scores of the different age groups. Therefore we may assume that age does not affect the scores on knowledge sharing.
Company size
As mentioned earlier, the company size is measured by means of the number of employees working within that company. As table 4.10 shows, a large part of the sample consisted of mangers working in small (less than 25 employees) companies. There are countless small family owned organisations in Java. Therefore is not strange that a large part of the sample consist of these small companies. Also it is plausible that managers from smaller companies are less reluctant to participate in a research such as this. The relatively large number of bigger companies may be due to the fact that the contact persons in Java are acquainted with numerous managers of companies this size.
Page | 36
Number of employees Frequency Per cent
<25 62 44.9
25-49 16 11.6
50-100 24 17.4
100< 36 26.1
Total 138 100
When testing the effect company size has on Humanness it shows that the means for the different size groups are very similar (appendix 5.2.3). The standard deviations however are quite diverse. The homogeneity of variance test results in significant differences for all four Humanness dimensions. This means that the assumption that the different size groups score variances are (roughly) equal is rejected. However, the ANOVA test does not show significant differences between the groups. Also the applied post hoc Scheffe test does not indicate significant differences scores. Therefore we may assume that company size does not significantly affect the scores on Humanness.
Sigger et al.’s (2010) test for the Tanzanian sample results in the same findings: there are no significant differences between the different size groups on the Humanness scores.
Next the effect of company size on knowledge sharing scores is assessed (appendix 5.2.4). The means for the different groups are quite similar. When looking at the standard deviation a somewhat larger difference can be found for the dimension leadership & corporate culture. The homogeneity of variance test supports this finding by showing a significant difference in variance for this dimension. The ANOVA test however does not show a significant difference in scores between the groups. Therefore we may assume that company size does not have a significant impact on Leadership &corporate culture scores.
The ANOVA test does indicate significant differences between the groups on the scores for
Information Technology. The impact of this difference can be measured by calculating the effect size. The effect size for company size on Information Technology is .1003, which can be seen as moderate (Field, 2009).
Again a post hoc Scheffe test is performed to see if underlying differences can be identified. As appendix 5.2.4 shows this is not the case.
We can conclude that, although the t-tests and ANOVA tests did find some differences influenced by company sector and size, the effect sizes of these differences are only moderate and specific,
Page | 37 knowledge sharing dimensions score between moderate and high. We may therefore assume that the factors gender, age, company size and sector do not have significant impact on the scores on Humanness and knowledge sharing. These same conclusions were drawn by Sigger et al.(2010) for the Tanzanian sample, which increases the comparability of the results.
Page | 38
Chapter 5: Findings in Indonesia
Now that is determined that the measurement scales have done their work properly and are not significantly influenced by the control variables, we can continue analysing the data. Before applying the more advanced statistic tests in order to test the hypotheses, we first look at the descriptive statistics of the data in order to get a first impression of the dimensions of Humanness and knowledge sharing. The means and standard deviations are calculated for all dimensions to assess the level of Humanness and knowledge sharing (see Section 5.1). After the level of Humanness and knowledge sharing is determined the relation between the both constructs is assessed by applying a correlation test (see Section 5.2). Finally, the relations found with the correlation test are assessed with a regression analysis, to indicate the level and direction of the found relations (see Section 5.3). These results will be compared with those from Tanzania which can function as our expectation level.
5.1. Descriptive statistics
In order to determine the presence of Humanness a three level classification is used (Sigger et al., 2010) (Table 5.1). For the determination of attitude towards knowledge sharing the following classification is used (Scholtens, 2011): [1] negative attitude; [2] moderate attitude; and [3] positive attitude (Table 5.1).
Table 5.1: Score classification
Score 2.4 or less 2.5-3.5 3.6 and above
Degree of Humanness low moderate high
Willingness towards knowledge sharing low moderate high
Page | 39 of 4, which is considerately higher than the range for other the dimensions (Appendix 6.2). More descriptive statistics can be found in appendices 6.1 and 6.2.
As Table 5.2 shows, the means and standard deviations for the Humanness dimensions from Tanzania and Indonesia are very similar. The Tanzanian sample scored slightly higher on solidarity and the Indonesian sample slightly higher on respect, but both differences are still rather small. Likewise, the differences in scores for the knowledge sharing dimensions are minimal between the two countries (Table 5.3). Tanzania has a somewhat higher score on Information Technology, but its standard deviation is exact the same as that found in Indonesia.
Table 5.2: Descriptive statistics for Humanness and dimensions
Mean Standard deviation
Tanzania Indonesia Tanzania Indonesia
Humanness 3.93 3.93 .46 .40
compassion 4.10 4.14 .48 .43
solidarity 3.77 3.54 .54 .55
survival 4.33 4.23 .49 .43
respect 3.51 3.82 .76 .54
Table 5.3: Descriptive statistics for knowledge sharing and dimensions
Mean Standard deviation
Tanzania Indonesia Tanzania Indonesia
Page | 40
5.2. The correlation between Humanness and knowledge sharing
At this time we have determined that Humanness as a management concept is indeed present in Java. Also knowledge sharing is widely prevailing. Next it is interesting to see whether the presence of Humanness is related to the attitude towards knowledge sharing. Therefore the first step is to determine whether correlations can be found between the two constructs and their dimensions. This will be done with the use of the Pearson correlation test, which results in the size and significance of the correlation given by the Pearson correlation coefficient (r). As described in the hypotheses section, all relations are divided into three levels (see figures 5.1-5.4). This section will discuss the correlation within these different levels one level at a time.
Figure 5.1: Level 1
Level 1: constructs Humanness and knowledge sharing
According to the Pearson correlation test, Humanness is significantly correlated to knowledge sharing with a score of .833 (appendix 7.1).
The Pearson correlation score was found to be significant at a .01 level (2-tailed) and with an N of 138 (Table 5.4). This is a high score because Pearson correlation coefficients do not get higher than 1.
Table 5.4: Correlation between Humanness and knowledge sharing Humanness
Knowledge sharing .833
Level 2a: Humanness dimensions and knowledge sharing
Next, it is also interesting to measure the correlation between the individual Humanness dimensions and knowledge sharing (appendix 7.2). As Table 5.5 shows, the Humanness dimensions have
Page | 41 Table 5.5: Correlations between knowledge sharing and dimensions of Humanness
Knowledge sharing Survival .700 Compassion .688 Solidarity .680 Respect/dignity .689 Figure 5.2: Level 2a
Level 2b: Humanness and knowledge sharing dimensions
Then the correlations between Humanness and the separate dimensions for knowledge sharing are calculated (Table 5.6). The Pearson scores range from .569 to .800 and are all significant at a level of .01(2-tailed) and an N of 138. The highest correlation can be found between Humanness and Leadership &corporate culture with a Pearson score of .800. These results imply that both knowledge sharing and its dimensions are correlated to Humanness.
Table 5.6: Correlations between Humanness and dimensions of knowledge sharing Humanness
Page | 42
Figure 5.3: Level 2b
Level 3: Humanness dimensions and knowledge sharing dimensions
Finally the correlations between all individual Humanness dimensions and all knowledge sharing dimensions are assessed. As Table 5.7 shows all expected relations are significantly correlated. All Pearson scores given in Table 5.7 have an N of 138 and have significantly level of .00 (2-tailed)
Table 5.7: Correlations between Humanness dimensions and knowledge sharing dimensions Compassion Respect Solidarity Survival
Page | 43 Figure 5.4: Level 3
The Pearson correlation tests showed that there are significant correlations between (the dimensions of) Humanness and (the dimensions of) knowledge sharing. In order to test the
Page | 44
5.3. Regression analysis
After determining that there are indeed correlations to be found between Humanness and knowledge sharing, the strength and direction of these correlations can be assessed. This will be done by applying multiple linear regression models. Again, this will be done according to the three levels, starting with the constructs Humanness with knowledge sharing and ending with each dimension of Humanness with each dimension of knowledge sharing. For each analysis will be checked whether the regression model is fit to use.
5.3.1. Regression analyses of three levels
Level 1: Humanness and knowledge sharing
First of all a regression analysis between the both constructs is performed, in which Humanness is identified as independent variable and knowledge sharing is identified as dependent variable (Appendix 8.1).
As Table 5.8 shows, the test results in a R² of .695, which means that 69.5 per cent of all variability in the attitude towards knowledge sharing can be explained by the variance in the presence of
Humanness.
The standardised beta (std. B) represents the number of standard deviations that the dependent variable will change of the dependent variable that would result from a change of one standard deviation of the independent variable. It shows the size of the change, but also the direction.
Page | 45 Table 5.8: Regression analysis level 1
Humanness
r R² Std. B Knowledge sharing .833 .695 .833**
Figure 5.5: Standardised beta level 1
Level 2a: Dimensions of Humanness and knowledge sharing
Next the relation between the dimensions of Humanness (independent variables) and knowledge sharing (dependent variable) will be measured (Appendix 8.2). Figure 5.6 shows the standardised beta and its significance for each hypothesis.
Compassion and knowledge sharing
Table 5.9 shows that R² has a value of .696, which means that the variance in four dimensions of Humanness together, explain 69.6 per cent of the variability in the attitude towards knowledge sharing. The standardised betas range from .210 to .271 and are all significant at a level of significance of .01. With a regression efficient of .271, respect is the most influential variable. Compassion is the least influential with a correlation efficient of .210. All standardised betas are significant at a level of .01. Because there are now several variables in the linear model (one dependent and four independent) the correlation coefficient (r) and the standardised beta (std. B) have different values. While the Pearson correlation coefficient (r) shows the strength of the interdependence of the variables (how well a linear model explains the relation between both variables), the standardised standardised betas (std. B) represents the actual change of the dependent variable for each independent variable.
Page | 46
Table 5.9: Regression analysis level 2a Knowledge sharing r R² Std. B Compassion .688 .696 .210 Solidarity .680 .263 Survival .700 .265 Respect .689 .271
Figure 5.6: Standardised betas level 2a
Level 2b: Humanness and dimensions of knowledge sharing
According to the regression analysis, the variance in the presence of Humanness can explain 43.3 per cent of the variability in employee motivation (Appendix 8.3) (Table 5.10). The standardised beta between the two variables is positive with a value of .7658 and is significant at a level of .01. These results support hypothesis 3a (Managers that score high on Humanness values will show a higher score on employee motivation than to those who do not score high Humanness values). Figure 5.7 shows the standardised beta and its significance for each hypothesis.
Page | 47 Appendix 8.5 shows that 32.4 per cent of the variability in Information Technology can be explained by the variance in the presence of Humanness. The correlation between the two variables is again positive with a value of .569. This correlation is significant at a significance level of .01. Therefore hypothesis 3c (i.e. Managers that score high on Humanness values will show a higher score on Information Technology than managers that score low on Humanness values) is supported.
Table 5.10: Regression analysis level 2b Humanness
r R² Std. B Employee motivation .658 .433 .658 Leadership & corporate culture .800 .639 .800 Information Technology .569 .324 .569
Figure 5.7: Standardised betas level 2b
Level 3: Humanness dimensions and knowledge sharing dimensions
Dimensions of Humanness and Employee motivation
Table 5.11 gives an overview of the regression analysis for the third level. Figure 5.8 shows the standardised beta and its significance for each hypothesis. The red lines indicate rejected
Page | 48
higher score on employee motivation than managers that score low on respect/dignity) is rejected. Hypotheses 4a, 4b and 4c are supported, which means that managers that score high on respectively compassion, solidarity and survival have higher scores on employee motivation compared to those who have low scores on the first variables.
Table 5.11: Regression analysis level 3
Employee motivation Leadership & corporate culture Information Technology
r R² Std. B r R² Std. B r R² Std. B Compassion .568** .470 .215* .645** .683 .161* .487** .350 .248* Solidarity .576** .296** .570** .041 .508** .263** Survival .602** .315** .739** .409** .375** -.069 Respect .445** -.026 .705** . 359** .497** .244* ** Significant at a level of .01 *Significant at a level of .05
Figure 5.8: Standardised betas level 3
-.026 Solidarity Leadership & corporate culture Compassion Survival Respect/dignity Employee motivation Information Technology .215* .296** .315** .041 .263** .409** .244* .161* .248* -.069 .359** H4d H5b H6c H4a H4b H4c H5a H6a H6b H5c H5d H6d
Dimensions of Humanness and Leadership & corporate culture
Leadership & corporate culture has the highest R² (Appendix 8.7). The variances in the four
dimensions of Humanness together explain 68.3 per cent of the variability in leadership & corporate culture.
