FACTORS INFLUENCING KNOWLEDGE SHARING

FACTORS INFLUENCING KNOWLEDGE SHARING IN IMMERSIVE VIRTUAL WORLDS:
AN EMPIRICAL STUDY WITH A SECOND LIFE GROUP

GRZEGORZ MAJEWSKI, DR ABEL USORO
School of Computing, University of the West of Scotland
Paisley, PA1 2BE, UK
grzegorz.majewski@uws.ac.uk, abel.usoro@uws.ac.uk


Immersive virtual worlds such as Second Life have recently gained much attention from education and business because of its adaptability to address real world challenges such as: on-line presentations, meetings, collaboration, 3D data visualization and on-line knowledge sharing. It is also possible to use immersive virtual worlds for the purpose of facilitating knowledge sharing in Virtual communities of practice (VCoP). Varieties of such communities exist in Second Life and help their members to achieve their goals. There is however not enough research into knowledge sharing in immersive virtual worlds. Therefore the purpose of this research is to fill this gap in knowledge. The conceptual model proposed by Usoro and Majewski (2008) will be applied to knowledge sharing in immersive virtual world environment. It will be investigated what factors are the most influential while residents of immersive virtual worlds share their knowledge. As a first attempt to validate the research model, quantitative and qualitative research was carried out with participants of Second Life group.

1. Introduction
Knowledge sharing is valued by modern organisations because they view their success to be highly dependent not only on their employees’ possession of knowledge but more importantly on the dynamic flow of this knowledge between the employees. Research so far indicates that Collaborative Virtual Environments (CVE) may be effective ways of capturing and sharing knowledge (Tomek, 2001). CVE is a “software environment that emulates some of the features of the real world” (Tomek 2001, pp 458-459). The characteristics of the real world that are usually emulated are: the concept of space, the representation of an object, the representation of a human (in a form of avatar), and various tools that can be used to interact with other objects.
A CVE “channels” knowledge sharing activities through software medium. However instead of a knowledge giver having to type into the system explanations of some accomplished task, he or she can easily perform the task in a CVE using his or her electronic surrogate (avatar). The playing out of activities or knowledge is very useful especially in the case of sharing tacit knowledge which may not be fully realized by even their owners. A CVE, such as Second Life (SL) essentially captures relevant parts of the work process, organizes it and provides data retrieval and data mining functions. It may motivate employees to share their knowledge by combining work and entertainment-like functions (Tomek 2001, p 459). The 3D environment can also “help develop a common understanding in a collaborative mind set and engage people through appealing and memorable experiences” (Schmeil & Eppler 2008, p 667). Apart from encouraging creativity at a high scale, the opportunity to interact with the 3D world via a virtual self – avatar - helps participants to develop much stronger social ties as compared with a static web discussion board or forum.
While employing SL in knowledge management and collaboration related activities it is necessary to consider the following characteristics of this environment (Tomek 2001, Ondrejka 2008, Schmeil & Eppler 2008):
• As a prerequisite for a successful groupware, physical topology is emulated as a common metaphor to give a natural perception of the environment.
• People, information and knowledge can be organized spatially.
• Awareness of co-workers, usage policies for tools and objects is enhanced.
• Content is produced by residents of the world; developers provide powerful tools designed to be used by everyone.
• Group and private chat functionality, as well as object sharing provide inherent collaboration possibilities are provided (additionally all communication can be logged instantly).
• Social/collaborative aspect – inherent collaboration between avatars – is facilitated.
• Constructivist aspect – playing or creating objects and so creating correlations and knowledge from current structures is inherent in SL (Antonacci, 2005).
• Collaborative problem solving engaging several avatars is supported.
Second Life and other environments based on Virtual Reality resemble the physical reality and thus it is easier for the participants to engage with them and usually such engagement is to a higher degree. It comes however at a price – 3D models are more difficult to use (require more manipulation of the user’s avatar) and may distract from the communication (Tomek 2001, p 461). In comparison to a “flat” graphical CVE a truly 3D environment makes it possible to interact with information that is dynamic and interactive.

2. Theoretical Background
As it was discussed before through the use of virtual immersive worlds it is possible to engage employees to a much higher level as compared with traditional means. Edirisingha et al. indicate that 3D environments such as Second Life promote socialisation among participants (learners) and that is a “key stage for learning in online environments” (p. 459). Learning may be perceived as reception of knowledge provided by other participants. Immersive virtual worlds have the potential to generate much greater sense of presence and belonging to a given community or subculture. These in turn may result in increased levels of knowledge sharing. The feeling of presence helps reduce the emotional and psychological distance between geographically dispersed participants. Collaboration is successful when the “learners feel comfortable with each other and when they are provided with opportunities for socialisation” (p. 460). Online socialisation is more than just getting to know each other. It extends to participants establishing their online identities, finding others with whom to interact online, developing trust and mutual respect, and working on common tasks online.
In some cases a shared 3D virtual environment may enable the creation of a virtual community of practice, and in fact extend into real-life collaboration (Jarmon & Sanchez, 2008). It is necessary to understand that usually in virtual worlds such as Second Life, the main experience of the user is the experience of a variety of social networks with different goals (e.g.: building, playing, buy, selling, working). Second Life in this view is more of a social experience than a simple game. It may be perceived as both a social and a technical system. Jarmon & Sanchez (2008) describe the community grounded in the framework of Communities of Practice (Wenger 1998) – The Educators Coop Island. The technical system was based on the spiral of knowledge by Nonaka and Takeuchi (1995), where participants have opportunities to participate in a spiral of sharing explicit and tacit knowledge. The research found out that “through their experiential learning, the residents are building on their part knowledge and expertise and are improvising a new community of practice in Second Life.” (p. 70). Authors conclude the virtual world environment creates opportunities to experience and understand issues that may be difficult to comprehend without actually “living” through them in the real life. Moreover participants began to rapidly expand their networking activity by using other networks outside Second Life (Google social software – joint documents and calendars).

3. Research Model
Knowledge sharing in immersive virtual worlds is a complex phenomenon. As it was discussed previously social perspective is one of the most important ones. The perception of a given community a person (in a form of avatar) participates in is central to knowledge sharing processes. For the purposes of this research it will be investigated how the perception of community affects two most common knowledge sharing activities: knowledge provision and knowledge seeking. The importance of social ties in knowledge sharing is also emphasized by Chiu et al. (2006). It can be argued that people coming to a community in Second Life “are not just seeking information or knowledge to solve problems but they also treat it as a place to meet other people, to seek support, friendship and a sense of belonging.” (Chiu et al 2009, p. 1874). Kankanhalli et al. (2005a) agree. These ideas are put into the following hypotheses:
• H1: Perception of community is positively associated with knowledge provision.
• H2: Perception of community is positively associated with knowledge reception.
Perception of community is itself influenced by such factors as trust and norm of reciprocity. Kankanhalli et al. (2005a) concluded that reciprocity is positively related to knowledge sharing. The hypotheses that can be derived are that:
• H3: The norm of reciprocity is positively associated with the perception of community while providing knowledge.
• H4: The norm of reciprocity is positively associated with the perception of community while seeking knowledge.
Hypothesis H3 states that the expectation of reciprocity (i.e. of being helped in the future by other members) influences the way the community is perceived. Participants are also more eager to receive knowledge from a community with an established reputation than from a new one they do not know yet, and consequently the reputation and perception of the community determines the way its members treat one another. In the case of H4 it is also supported by the view that “the exchange of intellectual capital can be facilitated by norms of collaboration and sharing”. The opinions of superiors and peers affect human decisions to use a given technology and to seek knowledge (Kankanhalli et al., 2005b, p. 1158).
Another important factor influencing knowledge sharing in VCoP is trust. It is possible to distinguish the following components of trust: competence (expertise), benevolence and integrity (Usoro et al., 2007; Willem & Buelens, 2009). The study by Usoro et al. (2007) investigates how these components affect knowledge provision in VCoP and conclude that there is a positive association. Trust may also be seen as a facilitator of social ties (Chow & Chan, 2008) and thus has a positive impact on knowledge sharing. “The level of social trust influences expectations of a colleague’s intention and behaviour. Organizational members are thus more likely to expect those who are trustworthy to share their knowledge” (Chow & Chan 2008, p. 460). On the other hand social ties positively influence mutual trust (Van den Hooff & Huysman, 2009). In other words it seems that trust is highly associated with the social aspect of VCoP. A study by Lin et al. (2009) indicates that norms of reciprocity influence trust, which in turn affect knowledge sharing self-efficacy, perceived relative advantage, perceived compatibility and knowledge sharing behaviour. Usoro et al. (2007) link some trust components with reciprocity, while indicating that it further influences the sense of community: “Where the sense of community is strong and benevolence is high, community members are more likely to perceive knowledge as a public good, owned and maintained by the community. Conversely, if one’s sense of community’s benevolence is low, expectations of future reciprocity may likewise be low, and knowledge sharing is unlikely to be fostered” (Usoro et al., 2007, p. 203).
Trust also affects knowledge reception. It is trust in the credibility, expertise and reputation of the contributor (knowledge source) that matters. “Perceived value from knowledge seeking depends on the contributor’s (source’s) expertise and credibility, while perceived expectation of value is determined by trust, obligation, and the contributor’s willingness to help” (Bock et al., 2006). It is possible to say that trust in on-line community affects the perception of the community on the knowledge receiver side. If the community is perceived as competent in answering queries it is expected that a knowledge seeker will keep posting his or her questions. Other trust components count as well, because a knowledge receiver needs to be able to rely on the knowledge provided by the community. Thus benevolence and integrity are also important. It is possible to formulate the following hypotheses based on these facts:
• H5: Trust is positively associated with the perception of community while providing knowledge.
• H6: Trust is positively associated with the perception of community while receiving knowledge.
• H7: The Norm of reciprocity is positively associated with trust while providing knowledge.
• H8: The Norm of reciprocity is positively associated with trust while receiving knowledge.
These hypotheses are utilized in the construction of a conceptual model that will embrace both roles that can be played by the participants of VCoP. A simplified version of such research model is presented on Figure 1:

Figure 1 Conceptual model of factors influencing knowledge sharing in VCoP in immersive virtual environment.

4. Research Methodology
The hypotheses were tested by way of qualitative and quantitative research with the Lucky Tribe group in Second Life. The group was carefully selected to reflect the Second Life environment. This huge 3-D immersive virtual world embraces profit-oriented organisations as well as people that seek different sorts of entertainment. The idea behind the group was to promote various vendors across the grid as well as generate traffic on their lands. It is however achieved through the activities that are very often entertaining in their nature. The economic analysis of the whole venture is beyond the scope of this paper; however it is possible to say that in this case knowledge sharing between participants has some economic aspects in the background. By participating in the community and sharing knowledge participants may be rewarded in a variety of ways.
Lucky Tribe group at the time of the distribution of the questionnaire had more than 1,480 members. Participants of this group perform a variety of functions in Second Life for which they are awarded the group own “currency” – kudos points. Furthermore there are tools and practices that enhance the strength of the social ties: dedicated Heads Up Display (HUD) showing achievements of the given member and also ability to check status (by showing the number of kudos points a given person has). The practices that enhance the social ties are special events or quests where members have to gather together to collaboratively do a given task. This community actually expands to other technologies (such as flicker and blogs). For successfully accomplishing a given activity members are rewarded kudos points and a special badge that is visible in their HUD. The proposed hypotheses were operationalised into a set of questions measuring each of the hypotheses (see Appendix I). Additionally a special set of interviews were developed to further validate the model. The group owner was contacted and asked for permission to do the research. This is a different approach to the Second Life social research as compared to Messinger et al. (2009) who distributed the questionnaire directly to the participants. After permission had been granted by the group owner a few members of the community were asked to evaluate the questionnaire. Nearly all the measurement items were based on literature or were operationalised from the conceptual model. A seven point Likert scale was used to measure the relative importance of each item. The feedback of the community members was the basis for the improvement of the questionnaire items which were thus pilot-tested.
An online version of the questionnaire was designed and a link to it distributed by way of group notices (special group mechanism available in Second Life), so that all of the members had the opportunity to access it. Additionally, after some time, a reminder was sent through a group notice and a group instant message announcement. These additional steps were taken due to technical problems with the delivery of group notices experienced by Second Life at that time. Participants of the on-line questionnaire were awarded kudos points, they received a specially designed “Research Assistant” badge visible on their HUD and the first 150 of them were paid 100 Linden Dollars (Second Life currency equivalent to about $0.40) .
The questionnaire required 15 to 20 minutes to complete. It is important to note that this approach provided very good results (good quality and number of responses). One of the important implications for the future researchers is that when carrying out such a study in Second Life it is best to combine two kinds of incentives. The first incentive should be the one that can be used by the participants within immersive virtual environment (in the case of this research it was the badge and the kudos points), while the other one may be an incentive that “translates” to real world gains (linden dollars in our case). Such configuration is important as it motivates both the participants that engage in immersive virtual environment for its own sake as well as those who aim to achieve real world benefits. The members of the group are truly global and come from a variety of countries and backgrounds. The interviews were distributed to the group owner, top members (with top 20 kudos points) and random members of the community.

5. Data Analysis
Questionnaires were distributed on the 13th of September 2009. At the time of writing a total of 161 responses were received. Incomplete surveys were removed and after data cleaning there was a total of 152 responses. In terms of demographic data the participants of the survey were mainly from the U.S.A. (63%), followed by U.K. 10% (some of them selected England - 2% and Scotland - 1%), Netherlands (6%), Germany (6%), Canada (6%), Australia (4%) and some other countries. It is worth to note that there were no participants from Asian countries, which is quite surprising given their huge population. Participants were mainly female (71%) with some fraction of males (24%) and 5% did not specify their gender. These results provide insight that immersive virtual worlds are quite popular with women as compared with the game-based industry where males dominate. This may also indicate that the social part of immersive virtual worlds plays an important part in attracting women.
As far as age of participants is concerned the leading group are people below 20 and 29 (more than 50 participants), followed by those in their thirties (almost 40 participants). It is worth to mention that these two groups are usually associated with the most active consumers. Therefore Second Life may be perceived as a good platform for marketing products and services from real world. The number of participants decreases with the age. It is also surprising that the people below 20 were not that well represented. A possible explanation is that there is a separate grid (Teen Grid) for people below 18 years old. Most of the questionnaire participants held College Certificate (33%), followed by BSc (21%) and Secondary School Certificate (19%), MSc (10%) and PhD (1%). These results are not very surprising given the high percentage of the participants from the U.S.A. where education usually finishes with college training. As far as profession is concerned 22% of the participants were engage in the Science, followed by Social Sciences (16%) and Arts (15%). Some of the participants (e.g. one working as an accountant) indicated in the comments that the scale did not provide enough choices. It may be necessary to adjust the scale in future studies. The results indicate that the participants where active across many professions with little dominance of Science .

6.1 Reliability and validity of measures
Data gathered from respondents were first assessed with reliability test which is a very important tool in positivist research because it indicates stability and the fact that the research is repeatable (Golafshani 2003, p 599). Reliability is a “statistical measure of how reproducible the survey instrument’s data are” (Litwin, 1995, p 6). The most common forms of reliability are: test-retest, alternate-form and internal consistency. For the purposes of this research internal consistency is the most appropriate form as the other two would involve asking the same respondents to retake the same questionnaire (test-retest) or take a changed questionnaire (alternate-form).
Internal consistency reliability applies to groups of items that are perceived to measure different aspects of the same concept. Internal consistency may be measured by determining the value of Cronbach’s coefficient alpha. It measures internal consistency reliability among a group of items combined to form a single scale. It examines the degree to which the items measure the same underlying concept (Pallant 2005, p 95). In other words it is possible to state that it mirrors the homogeneity of the scale – reflects how well various items harmonize with each other to measure different aspects of the same variable. It is usually expressed as a correlation coefficient and values of 0.70 and above are generally accepted as representing good reliability (Litwin 1995, pp 24-31). In the case of short scales (e.g. fewer than ten items) low Cronbach values may occur. This statistics when performed on the whole questionnaire returns value of 0.686 (0.708), which is slightly less than the goal of 0.7. It still may be considered as a good reliability. In other words the questionnaire items measure a very similar concept – in this case it could be the knowledge sharing in immersive virtual worlds. Figure 2 presents the reliability statistics.
Cronbach's Alpha Cronbach's Alpha Based on Standardized Items N of Items
.686 .708 21
Figure 2 Reliability statistics for the whole scale
Besides measuring the reliability of the scales it is also necessary to examine their validity. Validity is a term which describes how well a questionnaire or survey measures what it is meant to measure. For example an item that is supposed to measure pain should measure pain and not some related variable (e.g. anxiety) (Litwin, 1995, p 33). It gives the researchers, their peers and the scientific society the assurance that the methods chosen are appropriate for searching the truth (Usoro et al, 2007). There are different types of validity that are usually measured when evaluating the performance of a questionnaire: face, content, criterion and construct (Litwin, 1995, p 33).
Construct validity is the most valuable and most difficult way of evaluating a questionnaire. This is due to the fact that it measures how meaningful the scale or questionnaire is when in practical use. It thus requires years of experience (Litwin, 1995, p 33-43). It was not possible to employ this validity technique to the full given the limited time scope of the study. Instead a substitutive method of factorial validity was used. This is a different approach where statistical techniques are employed to explain the construct validity instead of a long lasting research. Factorial validity with the most common technique, such as principal components analysis or confirmatory factor analysis, has been identified as one of the most appropriate in IS research (Straub et al, 2004, p 10). The results of the principal components analysis are provided in Appendix II in supplementary files. Five factors were extracted. From the analysis it appears that the first factor is highly correlated with positive aspect of the constructs discussed before. It is an indirect indication that there is a significant relationship between the items. The second factor is highly correlated with questions that were added to balance the scale and avoid positive slant (except the questions that could have been misunderstood by the participants). These two factors explain almost 50% of the total variance. The other three factors are scattered among different constructs and therefore explanation of what they could mean is much more difficult. The principal components analysis provides rather moderate support of the hypotheses. Therefore in the next step the hypothesis testing will be based on the original variables and constructs.

6.2 Hypothesis Testing
In order to test the hypotheses mentioned bivariate correlation analysis was performed. The results are grouped into two perspectives: knowledge provision and knowledge seeking .
Results for the knowledge provision indicated that there is a variety of significant correlations. Items that are measuring reciprocity (first three questions) are significantly correlated with those that are measuring trust. Only one out of nine possible correlation turns to be insignificant (NR1 and TR2). They are also correlated with the variables measuring perception of community. NR3 was added to balance the scale and avoid the positive slant of the scale; and it is clearly negatively correlated with the other variables measuring reciprocity (NR1 and NR2) and with most of the items measuring trust (TR1 and TR3) and perception of community (PC1 and PC3). It is also significantly and positively correlated with items TR2 and PC2 that were supposed to measure the distrust and negative sides of the community (in a similar fashion to avoid the positive slant). The items TR2 and PC2 behave in a similar fashion (significant negative correlation within the scale and with those variables that measure the given phenomenon from a positive point of view). The data is therefore a strong evidence of the hypotheses H3 and H7.
The questions measuring trust indicate high correlation with the variables measuring perception of community. Only TR2 and PC3 seem not to have a significant correlation. Additionally questions that were added to balance the scale behave in a similar fashion to the ones presented previously. They have significant positive correlations with each other and highly negative correlations with those that measure the advantages of trust and perception of community. This is a strong evidence in support of hypothesis H5.
There is also a strong correlation between items measuring perception of community and questions on knowledge provision. All of the correlations are significant except PC2 and two variables - KP2 and KP3. Additionally, variables PC1 and PC3 measuring positive aspects of the community are strongly negatively correlated with question KP1 that was added to the knowledge provision scale to balance it and avoid a positive slant. Although the correlation is not as strong as in the previous cases it is still in strong support of hypothesis H1.
Norms of reciprocity play an important role in the knowledge seeking behaviour as well. Each question on reciprocity is highly positively correlated with those measuring trust in knowledge seeker perspective. However the items that were added to both scales in order to balance the scales and avoid a positive slant do not match that well with each other and with those measuring positive aspects. The first question on trust, although formulized in a negative way, seems to be positively correlated with the items measuring reciprocity from the positive perspective. It is still possible that there is a strong relationship between reciprocity and trust in knowledge reception perspective. However the exact nature of this relationship needs to be further studied. The relationship between reciprocity and perception of community is much clearer than that. It is even clearer than in the knowledge provision perspective. All the reciprocity items are significantly correlated with those measuring perception of community. Moreover the items that were added to avoid a positive slant (NR3 and PCKR3) are only positively correlated with each other, while highly negatively correlated with those measuring a positive perspective. Therefore hypotheses H4 and H8 are valid. Hypothesis H8 which addresses norms of reciprocity, trust and knowledge sharing needs further elaboration and investigation into the questions themselves.
Hypothesis H6 which is about the relationship between Trust and Perception of Community in knowledge seeker perspective has similar problem with Hypothesis H8 that refers to the norms of reciprocity and trust. All the items are significantly correlated; however the question measuring the negative aspect of trust (TKR1) seems to have a positive relationship with questions measuring positive aspects of the community. On the other hand the variable that was supposed to measure the positive aspects of trust (TKR3) is negatively correlated with the questions measuring the positive aspects of the community and positively correlated with PCKR3 that was measuring the negative aspects of the community. In this case it is necessary to get back to the questions and from the quick analysis it appears that questions 13 and 15 (TKR1 and TKR3) are very similar to each other and use similar words (“afraid” and “not afraid”). This could be the reason and could mislead the participants of the survey. It is advised that these questions should be reformulated in any future studies. It is still possible to claim that the hypothesis itself was correct.
There is no doubt that the data provided is in strong support of hypothesis H2. All the variables measuring perception of community are in significant relationship with the items measuring knowledge reception. Additionally questions added to balance the scale and avoid a positive slant (PCKR3 of perception of the community scale and KR3 of the knowledge reception scale) are positively correlated only with each other, while they maintain significant negative correlation with the rest.

6.3 Interviews
Qualitative research was done in parallel with the distribution of the questionnaire. Out of ten planned interviews only three took place. Respondents were from the top 20 list of participants with most kudos points. It may be an indication that they associated themselves with the community to a much higher degree and therefore were more eager to share their experiences. Answers to question one were in favour of the lucky tribe group. It was termed as being friendly and polite. It turned out that although these three persons were high achievers they were also eager to help other members. Two of them try to treat all the members the same. One of the interviewees indicated that it is normal that friends are closer to them than other members are. Two of them find their own ways to do things in most cases before the rest (thus their high scores). One interviewee tests the knowledge provided by others and based on it reasons on future occasions. It would be an indication that on average they play knowledge provision role to other group members as indicated by the words of one of them: “I tend to figure things out before most others in the group”.
They seem to share the same view and see the group as friendly companion with friendly competition. One of the interviewees said very clearly that the way the community functions (very well) “translates” from the contributions made to it. The difference was in the way the participants responded to the fifth questions. While one of them answered positively another one argued that it does not matter. The third one provided an insight into what happens when the group owner does not provide a central repository. In this case members themselves archive, store and share necessary information. They all have agreed that trust plays a very important role in the community life. The first interviewee indicated that trust is absolutely important and that reciprocity plays an important role in it as well. Another one extended this view to all other communities in Second Life. While answering the last question first, the interviewee indicated the problem that newcomers to the community may have difficulty locating whom to ask. Another participant answered that the group owner did a “great job with the ‘mentor tiki’, which encourages one member to help another”. Mentor tiki is a special ritual that establishes a somewhat formal relationship between two participants (one as a mentor and one as a student – similar to knowledge provider and knowledge seeker) and helps share knowledge within the community.

7. Findings
Both quantitative and qualitative data analyses indicate that the hypotheses tested were valid. An interesting point was that although Second Life was envisioned as a primarily entertainment platform it evolved (due to the effort of its participants) to provide environment where knowledge sharing activities do take place as well.
The 3D environments also provide opportunities for interaction and socialization unseen in any other environments. These in turn can result in higher levels of knowledge sharing among avatar which are electronic surrogates of real world individuals.
Table 1 summarizes the results and findings of both quantitative as well as qualitative study. The final outcome is based on the results of the quantitative and qualitative analysis. It is positive in all of the hypotheses tested, because there is a strong quantitative support in nearly every case (except H6 and H8). Additionally interviews are in strong support of the hypotheses mentioned.
Quantitative Analysis Qualitative Analysis Outcome
Hypothesis First Step Interview
H1 Strong Strong Positive
H2 Strong Strong Positive
H3 Strong Strong Positive
H4 Strong Strong Positive
H5 Strong Strong Positive
H6 Moderate Strong Positive
H7 Strong Strong Positive
H8 Moderate Strong Positive
Table 1 Summary of findings.
8. Conclusions and areas for further studies
Knowledge sharing in immersive virtual worlds may take much richer form than in a “flat” static web environment. It however still requires very similar set of factors to exist. Norms of reciprocity, trust and perception of community are among the most prominent ones. This research has endeavoured to build a research model with these factors as constructs with hypotheses to relate them to each other. The analysis of the questionnaire responses from the Second Life group as well as interviews indicate that majority of them were valid. The obvious next step is the testing of the model with a larger sample as well as with other groups in Second Life, which share knowledge in a different way (virtual universities, campuses, laboratories, etc).
The findings of this research may be important for the organizations involved in the knowledge sharing in Second Life or other immersive virtual worlds or for those that plan to engage in this activity. It seems that such environments greatly enhance the most important factors of knowledge sharing as presented previously. It is however important to have clear goals of such knowledge sharing. While it may be expected that the engagement of users is much greater, the environment itself and increased socialisation level should not pose distraction for the participants. Therefore Second Life may find very good uses in the case of knowledge sharing where social part (e.g. perception of community, trust, social network) plays the crucial role. However in other cases it should be considered in advance whether or not to apply the pure flat 2D web instead or a combination of both so as to avoid the distraction of social activities.
An additional general implication for the people thinking of carrying out, in Second Life, research that includes questionnaires and interviews is that if any incentives are provided it is good to combine the incentives that can be utilized inside the virtual world with those that can yield real world profit for the participants. By doing so it is easier to a cover wider spectrum of participants - those that are involved in immersive activities for their own sake only and those who also try to gain profit from these activities in order to enjoy the profits in real world.
Areas for further research emerging from this study are the analysis of the impact of economic aspects on knowledge sharing in immersive virtual worlds and further analysis of the factors mentioned (with reworking of some of the questionnaire items). In the first case economic impact may play an important role. It could be analyzed how a variety of incentives could encourage members to provide and seek knowledge. In the second case some of the factors could be further elaborated. For example different aspects of trust (as indicated by Usoro et. al., 2007) could be analyzed.

References
Antonacci, D. & Moderass, N., (2005), Second Life: The educational possibilities of a massively multiplayer virtual world (MMVW), Retrieved July 14 2009, from http://connect.educause.edu/library/abstract/SecondLifeTheEducati/43821
Bock, G-W., Kankanhalli, S. S. & Sharma, S., (2006), Are norms enough? The role of collaborative norms in promoting organizational knowledge seeking, European Journal of Information Systems, Vol. 15, 357 – 367.
Chiu, C-M., Hsu, M-H., Wang, E. T. G., (2006), Understanding knowledge sharing in virtual communities: An integration of social capital and social cognitive theories, Decision Support Systems, Vol. 42, 1872 – 1888
Chow, W. S. & Chan, L. S., (2008), Social network, social trust and shared goals in organizational knowledge sharing, Information & Management, Vol. 45, 458-465
Davenport, T. H. & Prusak, L., (1998), Working knowledge: How organizations manage what they know, Boston: Harvard Business School Press.
Edirisingha, P., Nie, M., Pluciennik, M. & Young, R., (2009), Socialisation for learning at a distance in a 3-D multi-user virtual environment, British Journal of Educational Technology, Vol. 40, No. 3, 458-479.
Golafshani, N., (2003), Understanding Reliability and Validity in Questionnaire Research, The Qualitative Report, Vol 8, No, 4 December, 597-607.
Hooff van den, B. & Huysman, M., (2009), Managing knowledge sharing: Emergent and engineering approaches, Information & Management, Vol. 46, 1-8
Jamaludin, A., Chee, Y. S. & Ho, C. M. L., (2009), Fostering argumentative knowledge construction through enactive role play in Second Life, Computers & Education, Vol. 53, 317-329
Jarmon, L. & Sanchez, J., (2008), The Educators Coop Experience in Second Life: A Model for Collaboration, Journal of the research centre for educational technology, Vol. 4, No. 2, 66-83.
Kankanhalli, A., Tan, B. C. Y. & Wei, K-K., (2005a), Contributing knowledge to electronic knowledge repositories: an empirical investigation, MIS Quarterly 29 (1), 113- 143.
Kankanhalli, A., Tan, B. C. Y. & Wei, K-K., (2005b), Understanding seeking from electronic knowledge repositories an empirical study, Journal of the American Society for Information, Science and Technology, Vol. 56 (11), 1156-1166.
Lin, M-J. J., Hung, S-W. & Chen, C-J., (2009), Fostering the determinants of knowledge sharing in professional virtual communities, Computers in Human Behavior, Vol. 25, 929-939.
Litwin, M. S., (1995), How to Measure Survey Reliability and Validity, London: Sage Publications.
Messinger, P. R., Stroulia, E., Lyons, K., Bone, M., Niu, R. H., Smirnov, K. & Perelgut, S., (2009), Virtual worlds – past, present, and future: New directions in social computing, Decision Support Systems, Vol. 47, 204-228.
Nonaka, I. & Takeuchi, H., (1995), The knowledge-creating company: How Japanese companies create the dynamics of innovation. New York: Oxford University Press.
Oliver, M. & Carr, D., (2009), Learning in virtual worlds: Using communities of practice to explain how people learn from play, British Journal of Educational Technology, Vol. 40, No. 3, 444-457
Ondrejka, C., (2008), Education Unleashed: Participatory Culture, Education, and Innovation in Second Life, The Ecology of Games: Connecting Youth, Games, and Learning – Digital Media and Learning, Cambridge, MA, The MIT Press
Pallant, J., (2005), SPSS Survival Manual, Maidenhead: Open University Press.
Schmeil, A. & Eppler, M. J., (2008), Knowledge sharing and collaborative learning in Second Life: A Classification of Virtual 3D Group Interaction Scripts, Journal of Universal Computer Science, Vol. 14, No. 3, 665-677.
Straub, D.; Boudreau, M. & Gefen, D., (2004), Validation guidelines for IS positivist research, Communications of the Association for Information Systems, Vol. 13, 380-427.
Tomek, J., (2001), Knowledge Management and Collaborative Virtual Environments, Journal of Universal Computer Science, Vol. 7, No. 6, 458-471.
Usoro, A., Sharratt, M. W., Tsui, E. & Shekhar S. (2007) Trust as an antecedent to knowledge sharing in virtual communities of practice. Knowledge Management Research & Practice, 5, 199-212.
Usoro, A. & Majewski, G. (2008), Trust and Risk as Critical Factors of Knowledge Sharing in Virtual Communities of Practice: A Conceptual View, Proceedings of the 9th European Conference on Knowledge Management, 443-452.
Wenger, E., (1998), Communities of Practice: Learning, Meaning and Identity, Cambridge. UK: Cambridge University Press.
Willem, A. & Buelens, M., (2009), Knowledge sharing in inter-unit cooperative episodes: The impact of organizational structure dimensions, International Journal of Information Management, Vol. 29, 151-160.

Appendix I (Operationalisation)
No Construct Operationalisation Source Questionaire item
1 Norms of reciprocity (NR) Reciprocity (KP) Chiu et al. (2006) p. 1874, Kankanhalli et al. (2005a), Lin et al. (2009), p. 932, Kankanhalli et al., 2005b, p. 1158, Usoro et al., 2007, p. 203, Gannon-Leary & Fontainha E, 2007, p. 6, Bock et al., 2006, p. 359, Chow & Chan, 2008, p. 459-460 Question 1
Question 2
Negative Aspect Question 3
2 Trust (KP) Kankanhalli et al. (2005a), Chiu et al. (2006) p. 1874, Usoro et al., 2007; Willem & Buelens, 2009, Chow & Chan, 2008, Van den Hooff & Huysman, 2009, Lin et al. (2009), p. 932, Bock et al., 2006, Gannon-Leary & Fontainha E, 2007, p. 6, Question 1 Question 3
Negative Aspect Question 2
3 Trust (KR) Bock et al., 2006, p. 358, Chow & Chan, 2008, Van den Hooff & Huysman, 2009, Lin et al. (2009), p. 932, Gannon-Leary & Fontainha E, 2007, p. 6, Question 2 Question 3
Negative Aspect (based on) Usoro et al., 2007 Question 1
4 Perception of community (PCKP) Kankanhalli et al. (2005a), Chiu et al. (2006) p. 1874, Chiu et al 2009, p. 1874, Zhang & Hiltz, 2003, Usoro et al., 2007, p. 203, Butler et al., 2002 p. 5, Gannon-Leary & Fontainha E, 2007, p. 6, Question 1 Question 3
Negative Aspect Question 2
5 Perception of community (PCKR) Bock et al., 2006, 2005b p. 359, Butler et al., 2002 p. 5, Gannon-Leary & Fontainha E, 2007, p. 6, Kankanhalli et al., 2005b, p. 1158, Usoro et al., 2007, Question 1 Question 2
Negative Aspect Question 3
6 Knowledge provision (KP) KP Frequency (based on) Usoro et al., 2007 Question 2
Question 3

Negative Aspect Chiu et al. (2006) p. 1874 Question 1
7 Knowledge reception (KR) KR Evaluation (based on) Usoro et al., 2007 Question 1 Question 2

Negative Aspect Question 3
Appendix II (Principal Component Analysis)
Component Component
1 2 3 4 5 1 2 3 4 5
NR1 .525 KP3 .332 -.675
NR2 .729 TKR1 .483
NR3 .775 TKR2 .760
TRP1 .789 .302 TKR3 -.682 .431
TRP2 .436 -.661 PCKR1 .424 -.364
TRP3 .727 PCKR2 .800
PCP1 .570 PCKR3 .377 .405
PCP2 .665 KR1 .646
PCP3 .545 -.363 KR2 .833
KP1 .810 KR3 .847
KP2 -.867
Extraction Method: Principal Component Analysis.
Rotation Method: Oblimin with Kaiser Normalization.
a Rotation converged in 19 iterations.