Knowledge Transfer and Collaboration in Distributed U.S.-Thai Teams

Knowledge Transfer and Collaboration in Distributed U.S.-Thai Teams

Saonee Sarker
Department of Information Systems
Washington State University

This article reports on a study that investigates factors influencing knowledge transfer in the context of cross-cultural distributed teams engaged in information systems development. The goal was to examine the validity of a four-factor framework of knowledge transfer (the "4 C Framework"), which proposes that capability, credibility, communication, and culture of the source significantly affects knowledge transfer. The framework is examined in the context of US-Thai distributed teams, as well as within the local subgroups. Results support the role of credibility and communication on knowledge transfer in the cross-cultural distributed teams, and within the local subgroups. Capability was not found to be related to knowledge transfer either in the distributed teams or within the local subgroups. Finally, culture of the source did affect knowledge transfer in the distributed teams, although in a direction opposite to that hypothesized.

Today, organizations' significant advancements in information and communication technologies (ICTs) and increased use of offshore outsourcing opportunities have resulted in different types of organizational work being conducted by formal, distributed, cross-cultural, and ICT-mediated teams (Carmel, 1999; Hossain & Wigand, 2004). The unending quest for ways to enhance organizational effectiveness has also prompted the increased use of informal semi-structured groups of individuals organized around a particular discipline or set of ideas (Lesser & Storck, 2004). Such groups are referred to as "communities of practice" and are seen to foster collaboration and help build relationships (Lesser & Storck, 2004).

Be it formal organizational work groups charged with the responsibility of completing a project within a specific duration, or be it a community of practice, knowledge sharing is critical to their existence and success (Kane, Argote, & Levine, 2005). As Levin, Cross, Abrams, and Lesser (2004, p. 36) argue, managers of such work groups are often faced with the question, "How can I encourage people to share what they know?" This sharing of knowledge becomes especially problematic when team members are distributed in time and space (Sarker & Sahay, 2004) and communicating only through an electronic medium (Cecez-Kecmanovic, 2001; Davenport & Prusak, 1998; Goodman & Darr,1996; Leonard & Sensiper, 2002).

In addition, since members of such groups frequently come from different geographical locations (and therefore have diverse educational and professional backgrounds and areas/levels of expertise), they often lack a shared frame of reference (Sarker & Sahay, 2003). This leads to a high "interpretive barrier" (Sole & Edmondson, 2002), which in turn affects the success of the collaboration (Goodman & Darr, 1996; Hossain & Wigand, 2004). If there is a successful transfer of knowledge among the individuals in the work group, they may create a shared frame of reference, making the investigation of factors inhibiting/enabling knowledge transfer in such contexts a critical topic for researchers and practitioners. Yet very little scholarly work has been done to investigate these inhibitors/enablers.

One prior study investigated knowledge transfer in U.S.-Norwegian virtual teams. That study proposed and tested a "4C Framework" outlining four critical source characteristics that enable knowledge transfer: capability, credibility, communication, and culture (Sarker, Sarker, Nicholson, & Joshi, 2003). A limitation of this study, however, was that the team members were drawn from U.S. and Norway, both western countries with similarities in several cultural dimensions (cf. Hofstede, 2001). Given that much of today's global collaboration (e.g., offshore outsourcing ventures) occurs between more distinct cultures (e.g., Western and Asian countries), it is important to explore how the factors identified by the 4C framework hold when the distributed teams consist of members from these more diverse cultures. This article's study builds on previous work (e.g., Sarker et al., 2003, forthcoming) and tests the 4C framework within distributed teams composed of members from two countries: the U.S. (i.e., having a "Western" culture) and Thailand (i.e., having an "Asian" culture).

Specifically, this study examines the influence of the four factors suggested by the 4C framework within the context of cross-cultural student teams engaged in a semester-long information systems development (ISD) project. ISD projects provide a suitable platform for examining knowledge transfer, given that such projects have often been described as a process of learning. ISD researchers argue that to successfully build a large and complex system, team members need to learn from each other continuously, and share knowledge regarding client requirements, capabilities of the new system, and the architecture of the computers (Curtis, Krasner, & Iscoe, 1988; Sarker et al., forthcoming).

The primary research question guiding this study is:

RQ 1: To what extent do the 4Cs proposed in the literature as having an effect on knowledge transfer in distributed teams consisting of "Western" (i.e., US and European) information system developers hold for teams where members are drawn from more distinct cultures (i.e., US and Asian)?

Further, noting that geographically-distributed teams often consist of local subgroups, a secondary research question is:

RQ 2: Do the same characteristics and behaviors identified by the 4C framework explain knowledge transfer within the local subgroups?

In the next section, I discuss the knowledge transfer literature, including its role in ISD, followed by a brief discussion of the 4C framework. Thereafter, I elaborate upon the research methods for testing the two research questions, and present the results. Finally, I provide a discussion of the study's results and contributions.

Transfer of knowledge from one set of individuals to another has been a key area of interest for knowledge management researchers. Alavi (2000) highlights the importance of knowledge transfer by suggesting that for superior performance of a social entity, knowledge generation and its successful transfer needs to take place. Cross, Parker, Prusak, and Borgatti (2004, p. 62) also posit the value of knowledge sharing in today's economy, "where collaboration and innovation are increasingly central to organizational effectiveness." Researchers on ISD teams also emphasize the importance of knowledge transfer among members (Carmel, 1999; Curtis et al., 1988). They argue that an ISD project involves activities that require the participation and contribution of all team members. To successfully build a large and complex system, team members need to learn continuously from each other regarding different issues, including the capabilities of the new system, application-specific algorithms, and the intentions of the customers as reflected in the requirements statements (Curtis et al., 1988). This transfer of knowledge is "often laborious, time consuming, and difficult" (Szulanski, 2000, p. 10), and it can become even more daunting in situations where knowledge is being transferred across time and space, such as in a globally distributed team (Alshawi & Al-Karaghouli, 2003; Davenport & Prusak, 1998; Sarker & Sahay, 200). Thus it is important to understand factors that impede or facilitate such transfer of knowledge.

Drawing on prior research, knowledge transfer in ISD is defined as the transfer of a source's ISD-related knowledge (both explicit knowledge such as technical know-how, and tacit knowledge pertaining to the management of large IS projects) to designated recipients (Sarker et al., forthcoming) within a cross-cultural distributed team engaged in information systems development. In the next section, the theoretical model is discussed in more detail.

Szulanski (2000) indicates that five basic elements can potentially influence the transfer of knowledge: channel, message, context, recipient, and source. Characteristics of the source have been identified as an important variable affecting knowledge transfer (Sarker et al., forthcoming). For example, studies of knowledge management have identified different characteristics of the source, such as their level of expertise, trustworthiness, etc., to be important "frictions" (i.e., inhibitors) of knowledge transfer (Davenport & Prusak, 1998; Hinds, Patterson, & Pfeffer, 2001; Szulanski, Cappetta, & Jensen, 2004). However, with the exception of the 4C framework, there is no known empirically testable system of propositions explaining the role of the source's characteristics in the extent of knowledge transfer. Therefore, in this study, the 4C Framework of knowledge transfer has been used for developing the research model.

The 4C Framework suggests that a source's capability with respect to his/her team members affects the extent to which he/she transfers knowledge. Capability of the source may be defined as the knowledge worker's hard skills (which in an ISD context could include technical know-how related to information systems development and their applications) and soft skills (which could include experience in working in large IS projects, communicating with users, etc.) (Davenport & Prusak, 1998). In addition, the 4C framework identifies the source's credibility, that is, his/her trustworthiness and reliability (Szulanski, 2000; Szulanski et al., 2004), his/her extent or quantity of communication with remote team members (Bresman, Birkinshaw, & Nobel, 1999), and his/her culture (Davenport & Prusak, 1998; Yoo & Torrey, 2002) as important factors affecting the extent of knowledge transferred by that individual. (See Figure 1 for a diagrammatic view of the model and Table 1 for the definition of the primary concepts used in this study.)

Figure 1. The 4C framework

Table 1. Definitions of the primary concepts in this study

Next, the effects of each of the above-mentioned factors on the extent of knowledge transfer in cross-cultural distributed teams are elaborated, and the research hypotheses are presented.

A critical factor enabling knowledge transfer in organizations is the presence of "smart people" (Davenport & Prusak, 1998, p. 88) who have the expertise necessary to accomplish the work. A source with a greater expertise than his/her remote members has the potential to transfer more knowledge to the recipients (Hinds et al., 2001; Zander & Kogut, 1994). Levin et al. (2004) argue that an individual who is perceived to be more knowledgeable (i.e., has more competence in a given subject area than a recipient) is likely to transfer knowledge to that recipient. The extent (or amount) of knowledge transferred is hypothesized to be proportional to the difference in knowledge levels of the source and the potential recipient.

In ISD-related projects, there are two arenas where expertise is particularly valuable: 1) technical skills, which include programming and database-related competence; and 2) analysis and project management skills, which include the ability to conduct requirements elicitation, relationship management, resource allocation and tracking, and application of systems development methodologies (Davenport & Prusak, 1998; Hoffer, George, & Valacich, 2002). In order to create a shared frame of reference within a distributed ISD team, such knowledge needs to be transferred to other members (Davenport & Prusak, 1998). Consistent with the views of von Krogh, Ichijo, & Nonaka (2000), who state that the "difference in knowledge expertise between corporate team of experts and local organizations drive knowledge transfer," it is argued that individuals who have higher levels of expertise than their remote team members would act as potent sources of knowledge transfer in a distributed team engaged in ISD.

Hypothesis 1: In a distributed ISD project, the ISD capability of an individual (with respect to his/her remote members) will positively affect the amount of knowledge transferred by that individual to his/her remote members.

Credibility of the source of information significantly affects knowledge transfer. When a source of knowledge is not perceived as credible, the advice and exemplars offered by the source are likely to be challenged and resisted (Walton, 1975), thereby reducing the extent of knowledge transfer. Credibility consists of the two related concepts of trust and reputation, which have been identified as important determinants of knowledge transfer (Yoo & Torrey, 2000). Trust is seen to improve "the quality of dialogue and discussion as a basis for organizational activities," thus facilitating the sharing of tacit knowledge (Ichijo, von Krogh, & Nonaka, 2000, p. 200). Along similar lines, Szulanski et al. (2004) contend that "trustworthiness of the source enhances … knowledge transfer." In addition, Levin et al. (2004, p. 37) argue that trust is the "magic ingredient" leading to knowledge transfer. When a source is perceived as untrustworthy, the recipient may consider the knowledge to be unreliable, and as a result, the recipient is less likely to internalize the knowledge communicated by the source (Szulanski, 1996).

On the other hand, Davenport and Prusak (1998, p. 101) believe that "reputation is a proxy for value that we use to evaluate the flood of information coming at us. We don't have time to look carefully at everything, so we select what we think will be worthwhile based on the reputation of the sender." Davenport and Prusak (1998, p. 101) further suggest that recipients of the information gauge the reputation of the sender by looking at the "performance" of that individual, rather than his/her hierarchical "status" in the organization. In a distributed team, performance becomes even more important, because 1) members often do not have a history of working together, and 2) given the flat structure of such teams, there is usually no permanent, universally-accepted hierarchical status of any individual (Saunders, 2000).

Hypothesis 2: In a distributed ISD project, the level of credibility of an individual perceived by remote members will positively affect the amount of knowledge transferred by that individual to his/her remote members.

Past research suggests that frequent communication assists in the creation of shared meaning and a common context within which the transfer process can be facilitated (Davenport & Prusak, 1998; Szulanski, 1996). Davenport and Prusak (1998, pp. 90-91) note that "in a knowledge-driven economy, talk is real work." They argue that it is through extended discussions that an individual's ideas, viewpoints, and beliefs are shared with, and made available to others. They further suggest that communication is the main mode by which workers "discover what they know," and "share it with their colleagues." This is even more true in ICT-mediated distributed teams, given that in such teams, computer-mediated communication forms the basis of all social action (Sarker & Sahay, 2003), including knowledge transfer (Venzin, von Krogh, & Roos, 2000). It is thus argued that an individual who engages in a higher extent (or amount) of communication will transfer more knowledge to his/her remote team members.

Hypothesis 3: In a distributed ISD project, the extent of communication between an individual and other remote members will positively affect the amount of knowledge transferred by that individual to those remote members.

Researchers consider culture to be another key determining factor in ICT-mediated distributed teamwork (Robey, Khoo, & Powers, 2000). In cross-cultural interaction (such as in a global virtual team), one of the primary factors that may affect the sharing of knowledge is the national culture of the source (Simonin, 1999;Yoo & Torrey, 2002). Cultural differences, often enacted in ISD teams as differences in attitudes towards system development, may also have some effect on knowledge transfer in an ISD context.

While culture has been studied in terms of a number of dimensions, there is growing acceptance of individualism (i.e., individualism-collectivism) being the key dimension for understanding differences in attitudes, values, norms, and behavior (Azevedo, Drost, & Mullen, 2002; Thomas, 2002; Triandis, 1995). The 4C framework draws on this dimension to explain the differences in knowledge transfer between cultures. Individualism/collectivism refers to the "relationship between the individual and the collectivity that prevails in a given society" (Hofstede, 2001, p. 209). It affects the way an individual thinks, and influences the way an individual processes, interprets, and perhaps even shares knowledge. Members from a more individualistic society view themselves as independent and can believe that "withholding information" is the key to success (Hofstede, 2001, p. 244). On the other hand, members from less individualistic societies believe that success of their collective unit depends on the ability to share knowledge with others. Thus, it is posited:

Hypothesis 4: In a distributed ISD project, members drawn from more collectivist cultures will tend to transfer more knowledge to remote team members.

In order to answer the primary research question, this study tested the 4C Framework of knowledge transfer using geographically (and culturally) distributed student teams working on a semester-long information systems development project. The use of student subjects has sometimes been criticized for a lack of generalizability. However, in this study, the longitudinal and intense nature of the project ensured that participants acted as professionals and not as typical students (Sarker & Sahay, 2003).

To address the second research question, the validity of the 4C framework was tested in the context of knowledge transfer among the local subgroups of the distributed teams. Given that this part of the study was exploratory in nature, the hypotheses were similar to the ones in the distributed team settings. In other words, the capability of an individual with respect to his/her local subgroup members, his/her credibility, and extent of communication as perceived by the local team members, were hypothesized to have a positive effect on his/her extent of knowledge transfer. Assuming that culture of the individual subgroup members would be fairly similar, the effect of culture was not hypothesized.

The sample for this study consisted of distributed teams working on information systems development (ISD) projects. In general, each team was comprised of usually four to five students enrolled in a systems analysis and design course in a U.S. public university, who were randomly teamed with about four to five students enrolled in a Thai university. Thus, each of the teams consisted of two local subgroups (one located in U.S. and one located in Thailand). There were a total of 11 teams with a useable sample size of 85, since analysis for this article was undertaken at an individual level.

The teams were required to develop computer-based application systems to solve business problems for organizations located in different parts of the world (e.g., the U.S., Hong Kong). For developing the computer-based applications, teams followed a systems development methodology, and were also responsible for creating the necessary documentation (e.g., system requirements statement, system designs, user manual, etc.)

The communication between the U.S. and Thai team members occurred primarily through the use of WebCT, an electronic communication tool that allowed online chats, document sharing, threaded discussion, and shared calendaring. Some teams also used other easily available tools, such as email and Instant Messenger, to supplement the capabilities of WebCT. Communication within the local subgroups occurred through face-to-face interaction and through other electronic means, primarily e-mail.

Data for this study were collected using online questionnaires. The questionnaires were administered to the distributed team members at two different points: before the start of the project (for measuring capability), and around the middle of the project (for measuring knowledge transfer, credibility, and extent of communication). In addition to the questionnaires, qualitative data in the form of chat transcripts and reflection documents created by each sub-group within the distributed team were also collected. While quantitative data were used for the analysis of the model, qualitative data were sometimes referred to for interpreting the results.

The dependent variable in this study was the extent of knowledge transferred by an individual in a distributed team to his/her remote team members or his/her local subgroup members. Following the suggestions of past literature (e.g., Darr & Kurtzberg, 2000; Davenport & Prusak, 1998; Sarker et al., forthcoming), knowledge transfer was measured by assessing the extent of learning of the recipient from a given source.

Specifically, the extent of ISD knowledge transferred by a source to remote recipients was measured by asking remote recipients about their extent of learning about ISD from that source. Two items were used to measure the extent of learning by the recipients: Team members were asked to specify, on a scale of 1 (Not at all) to 7 (To a great extent), the extent to which they had (1) learnt technical-ISD related skills and (2) managerial/behavioral ISD related skills from each of their remote team members. The extent of the knowledge transferred by an individual was the mean of the extent of learning of the remote team members from this individual (see Table 2 for the specific items).

Similarly, the extent of knowledge transferred by an individual to his/her local subgroup members was measured by the mean of the extent of learning of the local subgroup members from this individual.

This was measured as a difference between the capability level of the individual and that of his/her remote team members. First, the ISD Capability, comprising of technical ability and the IS project management ability of each team member, was measured (say, Capability 1) using a self-reported pre-questionnaire (validated and used in prior studies, see Sarker et al., 2005). The items measured a variety of different abilities ranging from knowledge of procedural programming to the ability to manage relationships between system development team members and users (see Table 2). Next, the mean technical and IS project management ability of all the remote team members for each individual team member (say, Capability 2) was computed. The Capability measure used in the analysis was the difference between Capability 1 and Capability 2.

For addressing the second research question, the capability of an individual with respect to his/her local subgroup members was computed. This was done by calculating the difference between the individual's capability and the mean capability of his/her local subgroup members.

Credibility was measured using the constructs of trust and performance. Each individual was asked to rate each of his/her remote team members on their trustworthiness and their performance in the project at that point of time (see Table 2). Thus, for a team of eight members (with four remote team members), each individual (whether located in Thailand or the U.S.) typically received four scores of trust and performance from his/her remote members. Based on these scores, a mean credibility score for each individual was computed.

Similarly, within the local subgroups, each individual was rated by each of his/her local subgroup members on trustworthiness and performance, from which a mean credibility score for each individual was computed.

Extent of communication, as perceived by remote members, was measured by asking each individual team member to specify the extent or amount of communication they have had with each remote team member, and a mean of this data for each individual participant was computed (see Table 2).

A similar technique was used for measuring the extent of communication within the local subgroups. Each individual was asked to rate each of his/her local subgroup members on their extent of communication, from which a mean score for each individual participant was computed.

Team member location was used as a surrogate for his/her culture (and hence for the degree of individualism of the team member). That is, culture for an individual was coded as 0 or 1 based on whether he/she was from the U.S. or Thailand. In a prior multinational study, Hofstede (2001) concluded that the U.S. has an individualism index value of 91 and ranks first among 53 countries on individualism, while Thailand has a score of 20 and ranks 39/41 on the individualism scale. Therefore, in this study individuals who were coded 0 on culture (i.e., from the U.S.) were seen to come from individualistic cultures, while those with a score of 1 (i.e., from Thailand) were seen to come from more collectivist cultures. The use of such an approach to measure culture, based on the published Individualism-Collectivism scores for countries (Hofstede, 2001) has been seen in other well-known studies such as those by Tan, Wei, Watson, & Walczuch (1998).

It is worth noting that a review of the data revealed that only 12 out of the 44 members from the U.S. originated from other cultures, with only five of them being residents of countries other than the U.S. (at the time of the study). On the other hand, all 41 members from Thailand originated and were residents of Thailand at the time of the study. From this it is reasonable to conclude that, overall, participants at each of the two locations (i.e., the U.S. and Thailand) represented the cultures of those countries.

Table 2. Questionnaire items

Since pre-validated instruments were used for measuring technical capability, IS-project management capability, knowledge transfer, and credibility, a confirmatory factor analysis of the items used for measuring the above-mentioned constructs was conducted, and the reliabilities were calculated to ensure the validity of the scale (see Table 3 for the reliabilities). Results indicated a good fit of the model with the data (see Table 4), with all items loading on their relevant factors at p<.001. A second-order factor analysis was also conducted to ensure that both technical ability and IS project management ability indeed loaded on the construct of ISD Capability. Results indicated that they both loaded significantly on the ISD capability construct (See Table 5).

Table 3. Reliability of the constructs

Table 4. First-order confirmatory factor analysis results

Table 5. Second-order confirmatory factor analysis for capability
^a - Overall model fit: Χ² (34) = 56.8, p < .10; GFI= .87; NFI= .92; CFI= .97; IFI= .96, RMSEA= .09

Next, a multiple regression analysis was used to assess the effect of Capability with respect to remote team members, Credibility, Communication, and Culture on knowledge transfer. (See Table 6 for the descriptive statistics and Table 7 for a summary of the result.)

Hypothesis 1 suggested that individuals with higher capabilities with respect to their remote team members would transfer more knowledge. Results did not support the hypothesis (H1: b = -.042, p = .150). While this result is clearly contrary to what was hypothesized, it is possible that highly-skilled members were conveying their knowledge in forms that were incomprehensible to the less-skilled team members, thereby reducing the absorption of that knowledge and learning by that recipient. Hypothesis 2 suggested that in a cross-cultural ICT-mediated distributed ISD project team, individuals who have high credibility will transfer more knowledge to their remote team members (H2: b = .563, p = .000). Results supported this hypothesis. There was also support for Hypothesis 3, which suggested that individuals who engaged in a high extent of communication with their remote team members transferred more knowledge to their remote counterparts (H3a: b = .210, p = .033). Finally, culture had a significant effect on knowledge transfer (H4: b = -.549, p = .001). However, results were in a direction opposite to the one hypothesized. In other words, individuals from more individualistic cultures (e.g., members from the U.S.) were seen to transfer more knowledge than those from the more collectivist cultures (e.g., members from Thailand). This could have been due to variations in computer-mediated communication styles across cultures (Kim & Bonk, 2002). Prior studies indicate that individuals from the U.S. tend to be more expressive and communicate more in online forums than their Asian counterparts. In addition, members from Thailand, even the communicative ones, owing to their lower language (English) competence, were not seen as transferring significant knowledge to their U.S. counterparts. The discussion section of this article provides an elaboration of this result.

Table 6. Descriptive statistics of constructs in the model U.S.-Thai distributed team

Table 7. Regression analysis testing 4C framework in the U.S.-Thai distributed team
^a - Dependent Variable- Knowledge Transfer (R² =.887; Adjusted R² = .881)
*** - < .01
** - < .05

Similar to the first analysis, a linear regression was utilized to test the role of the 4Cs within the local subgroups. The data were analyzed separately for U.S.- and Thai-based teams.

For the U.S. subgroups, capability difference did not have a significant effect on the extent of knowledge transfer (b= .016, p= .421). Credibility of an individual, and the extent of communication that he/she engaged in, had a positive effect on their extent of knowledge transfer (b=.578, p= .003; b= .318, p= .035).

For the Thai subgroups, capability difference and extent of communication did not play any significant role in explaining knowledge transfer (b= -.060, p= .250; b= .139, p= .185). However, credibility of an individual significantly affected knowledge transfer (b= .497, p= .000). (See Table 8 for the descriptive statistics and Table 9 for a summary of the results).

Table 8. Descriptive statistics of constructs in the model for the U.S.-Thai local subgroups

Table 9. Comparison of the test of the 4c framework within the U.S.-Thai local subgroups

^b - Dependent Variable- Knowledge Transfer (R² =.840; Adjusted R² = .831)
^c - Dependent Variable- Knowledge Transfer (R² =.627; Adjusted R² = .591)
*** - < .01; ** - < .05

To summarize, results from this study indicate that, for an individual to be perceived as an effective knowledge transferor by remotely-located recipients in a cross-cultural ICT-mediated distributed team engaged in an ISD project, he/she should extensively participate in electronic conversations, as indicated by the extent of communication, and be perceived as credible due to trustworthy behaviors and high performance. These results are consistent with a similar prior study involving U.S.-Norwegian distributed teams engaged in information systems development.

Contrary to expectations, differences in capability did not seem to enhance the extent of knowledge transfer in either the distributed or the local groups. While this may appear counterintuitive, the consistency of this finding across different types of distributed teams involving different cultures (e.g., the U.S. and Norway) suggests that the more knowledgeable and skilled members may be conveying their knowledge in forms that are incomprehensible to a less experienced or less skilled team member, thereby reducing the absorption of the knowledge and learning by the recipient, and reducing the overall extent of knowledge transferred. Even though this study used student subjects (who are often similar in dimensions such as capability), the differences in capability among the team members ranged from -4.70 to +4.46 (see Table 6). This suggests that there were members who had much higher ISD capability than their other members. It could have been possible that these members were not able to convey their knowledge successfully to their less knowledgeable team members.

Hinds et al. (2001) make a similar conclusion from their study of expert and novice behaviors in an electronic circuit wiring task. They argued that while "experts should have been well-positioned to convey their superior knowledge skills to novices, the organization of that knowledge, and particularly its level of abstraction, may make it difficult for them to do so" (Hinds et al., 2001, p. 1232). Another possibility is that knowledge of highly skilled individuals or experts is embodied in their actions, and it may be very difficult for the intended recipient to separate the knowledge from their actions (Davenport & Prusak, 1998). Swap, Leonard, Shields, and Abrams (2004, p. 182) refer to the concept of "pattern recognition" and argue that "the way in which experts exercise their knowledge is by calling on … experiences in a great variety of contexts to recognize patterns." This pattern recognition process tends to draw upon the "tacit dimensions" of the expert's knowledge that is extremely contextualized and therefore very hard to transfer (Swap et al., 2004, p. 183).

As expected, credibility and extent of communication both emerged as critical determinants of the extent of knowledge transferred. Individuals who were perceived to be trustworthy and high performers (i.e., had high credibility) and engaged in a high amount of communication with their remote team members were viewed to be significant knowledge transferors by their remote members.

As in the case of U.S.-Norwegian teams, culture played a significant role on the extent of knowledge transferred in this study of U.S.-Thai teams; however, contrary to expectations, in this study, members of individualistic cultures were viewed as transferring more knowledge than those in collectivist cultures. This clearly runs counter to prevailing thought on knowledge-sharing behavior. In order to make sense of this anomaly, some of the other qualitative data that were gathered during this semester-long project were used. These included chat transcripts and reflection documents created by each of the subgroups within the distributed teams at the end of the project, in which they discussed their experiences in the distributed ISD project, barriers they faced, how they attempted to overcome those barriers, and the lessons they learned through the process. A point that was repeatedly made by many of the U.S. subgroups was the limited and somewhat ineffective communication received from the Thai team members. This suggests that even if Thai members attempted to share knowledge through communication, they were not effective in getting through to the U.S. members, and thus, were not able to transfer a significant amount of knowledge.

One of the reasons for the lack of effective communication may have been the language barrier that existed among the subgroups. English was adopted as the official language for the project, and the language in which all of the project deliverables were going to be created. However, given that this was not the primary language of the Thai participants (at least at this university), this created a divide between the two remote subgroups within the distributed teams. The lack of language skills led to difficulty during chat sessions and instant messaging. The problem related to language and communication became such a constraint on the participants that some U.S. subgroup members sought the help of Thai students residing in the U.S. to act as interpreters during their electronic conversations. Similarly, Thai participants also sought the help of other individuals in Thailand who were fluent in English (who were not participants of the project, or who were members of a different group) to act as mediators between the U.S. and Thai members of their teams.

This problem related to communication and language helps make sense of the unexpected effect of culture (individualism/collectivism) in this study. It may be argued that Thai participants' lack of knowledge transfer was not related to their cultural values. Given that they come from significantly more collectivist cultures, sharing of knowledge was clearly valued by them. However, inhibitions related to their language skills perhaps made them shy away from engaging in extensive communication about new and difficult concepts with their remote participants. This argument is consistent with prior studies that have suggested that individuals who lack proficiency in a language often feel anxious about participating during collaborations (Kim & Bonk, 2002). In another study involving Danish and American college students, Bannon (1995) concluded that lack of fluency in English prompted the Danish students to stay away from computer-mediated communication. Davenport and Prusak (1998, p. 98) have also argued that "people cannot share knowledge if they do not speak a common language." In the case of the Thai students, it could have been possible that whatever they may have shared, due to their lack of fluency in the language, it was incomprehensible to their remote participants, leading to their (i.e., the remote team members') perception of little knowledge being transferred and lack of learning.

This provides some valuable lessons for fostering knowledge transfer in cross-cultural distributed teamwork. As the case of the Thai participants in the study highlights, intention to transfer knowledge is not sufficient to achieve success in knowledge transfer within a team. This problem multiplies when the team is composed of members from across the globe and communicating only through the electronic media, where text is the primary mode of sharing knowledge. Even if cultural values within the members foster such sharing, other factors, such as language skills (written and spoken), can act as a significant barrier to such transfer. Managers coordinating such teamwork need to be aware of this and take necessary steps to ensure that knowledge is successfully shared (e.g., selecting members who are fluent in the official language of the project, providing interpreters, etc.).

Within the U.S. subgroup, results seemed to be fairly consistent with findings for remote team members. The U.S. members seemed to value both credibility and the extent of communication of their local sub group members, and individuals with high credibility, and those who engaged in a higher volume of communication, were seen as transferring more knowledge. Again, capability difference did not have any effect on the knowledge transfer.

Results of the analysis of Thai subgroups were somewhat different. Within the Thai subgroups, individuals with high credibility were viewed as those transferring more knowledge. Interestingly, communication did not have any effect on knowledge transfer, suggesting that Thai subgroup members who were engaging in high volume of communication were not viewed as those transferring more knowledge. This result may be interpreted in terms of the cultural dimensions of high and low-context communication cultures. Most Asian nations (including Thailand) believe in high-context communication (Kim & Bonk, 2002, p. 24), which emphasizes "how intention and meaning can be best conveyed through the context (e.g., social roles, positions, etc.) and nonverbal channels (e.g., pauses, silence)." In other words, individuals from such cultures tend to value those who are likely to engage in implicit and reserved communication. Extending this line of reasoning, one might argue that individuals who engaged in a high volume of communication during the collaboration (i.e., were significantly less reserved) were not valued, and not perceived to be high knowledge transferors. In addition, the poorer statistical fit of the 4C Framework with the Thai subgroup data (as opposed to the U.S. subgroup data) also indicates that other factors (not captured by the 4Cs) may have been affecting knowledge transfer within the Thai subgroups.

Despite some limitations, such as the use of student subjects, this article makes a number of contributions, and in a number of ways continues to build on past research on knowledge transfer in cross-cultural, ICT-mediated distributed teams. First, it establishes the importance of communication, credibility, and culture of the source (identified by the 4C Framework) for knowledge transfer in the context of ISD teams with members drawn from countries with two dissimilar cultures. Further, the study shows some validity of the factors identified by the 4C framework within the local subgroups of a distributed ISD team, thereby highlighting the characteristics and behaviors that are important in situations where the source and the recipient are collocated, and from the same culture.

This study also provides some interesting lessons for distributed teamwork and cross-cultural collaboration involving members from Western and Asian countries. It suggests that important issues, such as the intention to transfer and share knowledge (as seen in collectivist) cultures, are not enough for fostering knowledge transfer in a distributed setting, and that other factors, such as language skills and communication competence, may play a critical role. In addition, the study highlights that the communication style preferred by cultures (high-context vs. low-context) may have significant impact on who is viewed as a knowledge transferor within a collaborative group.

While this study focuses on cross-cultural ISD teams, the results may, to some degree, be generalized to other group settings. For example, similar empirical patterns may be expected in virtual teams, where members are physically distributed, may share little or no history of working together, and may work on a variety of tasks (not ISD-related) for a given period of time (Wong & Burton, 2000). However, it must be noted that one would need to modify the operationalization of capability in order to study such non-ISD teams.

Results from this study may also be applicable to informal groups, such as a community of practice, which like distributed ISD teams, may consist of physically and culturally distributed members. Lesser and Storck (2004) argue that one of the primary differences between traditional teams and communities of practice is the fact that authority relationships within a team are often determined by the organization, while authority relationships within a community of practice usually emerges through interaction around expertise. This difference between a "team" (e.g., a cross-cultural ISD team) and a community of practice could have interesting implications for the model of this study. While in this study, experts were not seen as transferring more knowledge within an ISD team, in a study involving communities of practice (where expertise plays a more important role), differences in capability may have a more potent effect on knowledge transfer.

Overall, this study contributes to an understanding of knowledge transfer in cross-cultural collaboration, a topic that has been largely ignored by both collaboration and knowledge management researchers. Future research should focus on further examining the role of culture in knowledge transfer. Contrary to prior belief, this study showed that members of more individualistic cultures transferred/shared more knowledge. While qualitative data collected from the study helped to make sense of this anomaly in the current context, further research needs to be conducted before the provisional explanation offered in this article can be considered generalizable.

Research should also attempt to investigate the specific factors affecting knowledge transfer in Asian cultures (e.g., Thailand). Results from this study indicated that only credibility of the source plays an important role on knowledge transfer in such cultures. However, the lower statistical fit of the model for the Asian (Thai) culture suggests that other factors might play a role, and this warrants further investigation. Finally, future research should focus on examining what this study ignores: the effect of other group context factors (such as intra-group conflict, cohesion, etc.) on the extent of knowledge transfer. Clearly much work remains to be done in order to establish a coherent body of knowledge on this critical topic.

This research was supported by a grant from the International Business Institute, Washington State University.

Alavi, M. (2000). Managing organizational knowledge. In R.W. Zmud (Ed.), Framing the Domain of IT Management: Projecting the Future Through the Past (pp. 15-28). Cincinnati, OH: PinnFlex Education Resources Inc.

Alshawi, S., & Al-karaghouli, W. (2003). Managing knowledge in business requirements identification. Logistics Information Management, 16 (5), 341-349.

Azevedo, A., Drost, E. A., & Mullen, M. R. (2002). Individualism and collectivism: Toward a strategy for testing measurement equivalence across culturally diverse groups. Cross Cultural Management, 9 (1), 19-29.

Bannon, L. J. (1995). Issues in computer-supported collaborative learning. In C. O'Malley (Ed.), Computer Supported Collaborative Learning (pp.267-281). Berlin: Springer-Verlag.

Bresman, H., Birkinshaw, J., & Nobel, R. (1999). Knowledge transfer in international acquisitions. Journal of International Business Studies, 30 (3), 439-462.

Carmel, E. (1999). Global Software Teams. Upper Saddle River, NJ: Prentice Hall.

Cecez-Kecmanovic, D. (2001). What enables and prevents knowledge-sharing via computer-mediated communications? Journal of Systems and Information Technology, 5 (1), 115-134.

Cross, R., Parker, A., Prusak, L., & Borgatti, S. P. (2004). Knowing what we know: Supporting knowledge creation and sharing in social networks. In E. Lesser & L. Prusak (Eds.), Creating Value with Knowledge (pp. 61-81). Oxford: Oxford University Press.

Curtis, B., Krasner, H., & Iscoe, N. (1988). A field study of the software design process for large systems. Communications of the ACM, 31 (11), 1268-1287.

Darr, E. D., & Kurtzberg, T. R. (2000). An investigation of partner similarity dimensions on knowledge transfer. Organizational Behavior and Human Decision Processes, 82 (1), 28-44.

Davenport, T. H., & Prusak, L. (1998). Working Knowledge: How Organizations Manage What They Know. Boston: Harvard Business School Press.

Goodman, P. S., & Darr, E. D. (1996). Computer-aided systems for organizational learning. In C. L. Cooper & D. M. Rousseau (Eds.), Trends in Organizational Behavior, Vol. 3 (pp. 81-97). New York: John Wiley and Sons Ltd.

Hinds, P. J., Patterson, M., & Pfeffer, J. (2001). Bothered by abstraction: The effect of expertise on knowledge transfer and subsequent novice performance. Journal of Applied Psychology, 86 (6), 1232-1243.

Hoffer, J. A., George, J. F., & Valacich, J S. (2002). Modern Systems Analysis and Design. Upper Saddle River, NJ: Prentice Hall.

Hofstede, G. (2001). Cultures Consequences: Comparing Values, Behaviors, Institutions, and Organizations Across Nations. 2nd ed. Thousand Oaks, CA: Sage Publications.

Hossain, L., & Wigand, R. T. (2004). ICT enabled virtual collaboration through trust. Journal of Computer-Mediated Communication, 10 (1). Retrieved June 19, 2005 from http://jcmc.indiana.edu/vol10/issue1/hossain_wigand.html

Ichijo, K., von Krogh, G., & Nonaka, I. (2000). Knowledge enablers. In G. von Krogh, J. Roos, & D. Kleine (Eds.), Knowing in Firms: Understanding, Managing, and Measuring Knowledge (pp. 173-203). Thousand Oaks, CA: Sage Publications.

Kane, A. M., Argote, L., & Levine, J. M. (2005). Knowledge transfer between groups via personnel rotation: Effects of social identity and knowledge quality. Organizational Behavior and Human Decision Processes, 96 (1), 56-71.

Kim, K. J., & Bonk, C. J. (2002). Cross-cultural comparisons of online collaboration. Journal of Computer-Mediated Communication, 8 (1). Retrieved June 19, 2005 from http://jcmc.indiana.edu/vol8/issue1/kimandbonk.html

Leonard, D., & Sensiper, S. (2002). The role of tacit knowledge in group innovation. In C. W. Choo & N. Bontis (Eds.), The Strategic Management of Intellectual Capital and Organizational Knowledge (pp. 485-499). New York: Oxford University Press.

Lesser, E. L., & Storck, J. (2004). Communities of practice and organizational performance. In E. Lesser & L. Prusak (Eds.), Creating Value with Knowledge (pp. 107-123). Oxford: Oxford University Press.

Levin, D. Z., Cross, R., Abrams, L. C., & Lesser, E. L. (2004). Trust and knowledge sharing: A critical combination. In E. Lesser & L. Prusak (Eds.), Creating Value with Knowledge (pp. 36-41). Oxford: Oxford University Press.

Robey, D., Khoo, H., & Powers, C. (2000). Situated learning in cross-functional virtual teams. IEEE Transactions on Professional Communication, 43 (1), 51-66.

Sarker, S., & Sahay, S. (2003). Understanding virtual team development: An interpretive study. Journal of the AIS, 4 (1), 1-38.

Sarker, S., & Sahay, S. (2004). Implications of space and time for distributed work: An interpretive study of US-Norwegian systems development teams. European Journal of Information Systems, 13 (1), 3-20.

Sarker, S., Sarker, S., Nicholson, D., & Joshi, K. D. (2003). Knowledge transfer in virtual information systems development teams: An empirical examination of key enablers. Proceedings of the 36th Hawaii International Conference on System Sciences. Retrieved June 19, 2005 from http://csdl.computer.org/comp/proceedings/hicss/2003/1874/04/187440119a.pdf

Sarker, S., Sarker, S., Nicholson, D., & Joshi, K. D. (forthcoming). Knowledge transfer in virtual systems development teams: An exploratory study of four key enablers. IEEE Transactions on Professional Communication.

Saunders, C. S. (2000). Virtual teams: Piecing together the puzzle. In R. W. Zmud (Ed.), Framing the domain of IT management: Projecting the future through the past (pp. 29-50), Cincinnati, OH: PinnFlex Education Resources, Inc.

Simonin, B. (1999). Transfer of marketing know-how in international strategic alliances: An empirical investigation of the role and antecedents of knowledge ambiguity. Journal of International Business Studies, 30 (3), 463-490.

Sole, D., & Edmondson, A. (2002). Bridging knowledge gaps: Learning in geographically dispersed cross-functional development teams. In C. W. Choo & N. Bontis (Eds.), The Strategic Management of Intellectual Capital and Organizational Knowledge (pp. 587-604). New York: Oxford University Press.

Swap, W., Leonard, D., Shields, M., & Abrams, L. C. (2004). Using mentoring and storytelling to transfer knowledge in the workplace. In E. Lesser & L. Prusak (Eds.), Creating Value with Knowledge (pp. 181-200). Oxford: Oxford University Press.

Szulanski, G. (1996). Exploring internal stickiness: Impediments to the transfer of best practice within the firm. Strategic Management Journal, 17, 27-43.

Szulanski, G. (2000). The process of knowledge transfer: A diachronic analysis of stickiness. Organization Behavior and Human Decision Processes, 82 (1), 9-27.

Szulanski, G., Cappetta, R., & Jensen, R. J. (2004). When and how trustworthiness matters: Knowledge transfer and the moderating effect of causal ambiguity. Organization Science, 15 (5), 600-613.

Tan, B. C. Y., Wei, K.-K., Watson, R. T., & Walczuch, R. M. (1998). Reducing status effects with computer-mediated communication: Evidence from two distinct national cultures. Journal of Management Information Systems, 15 (1), 119-141.

Thomas, D. C. (2002). Essentials of International Management: A Cross-Cultural Perspective. Thousand Oaks, CA: Sage Publications.

Triandis, H. C. (1995). Individualism and Collectivism. Boulder, CO: Westview Press.

Venzin, M., von Krogh, G., & Roos, J. (2000). Future research into knowledge management. In G. von Krogh, J. Roos, & D. Kleine (Eds.), Knowing in firms: Understanding, managing, and measuring knowledge (pp. 26-66), Thousand Oaks, CA: Sage Publications.

von Krogh, G., Ichijo, K., & Nonaka, I. (2000). Enabling Knowledge Creation. New York: Oxford University Press.

Walton, R. E. (1975). The diffusion of new work structures: Explaining why success didn't take. Organization Dynamics, 3 (3), 3-21.

Wong, S., & Burton, R. M. (2000). Virtual teams: What are their characteristics, and impact on team performance? Computational and Mathematical Organization Theory, 6 (4), 339-360.

Yoo, Y., & Torrey, B. (2002). National culture and knowledge management in a global learning organization. In C. W. Choo & N. Bontis (Eds.), The Strategic Management of Intellectual Capital and Organizational Knowledge (pp. 421-435). New York: Oxford University Press.

Zander, U., & Kogut, B. (1994). Knowledge and the speed of the transfer and imitation of organizational capabilities: An empirical test. Organization Science, 6 (1), 76-92.

Saonee Sarker is an Assistant Professor of Information Systems at Washington State University. Her research interests include virtual teams and computer-mediated groups and group's adoption of technology. Specifically, her research examines knowledge sharing, leader emergence, and technological issues in systems development groups (both collocated and distributed).
Address: Department of Information Systems, College of Business and Economics, Washington State University, Pullman, WA 99164-4729 USA