International
Journal of
Educational
Technology

January, 2001 - Feature

The Construction of Shared Knowledge in an Internet-based Shared Environment for Expeditions (iExpeditions)

Minjuan Wang, San Diego State University;
James Laffey, University of Missouri-Columbia;
Melissa J. Poole, University of Missouri-Columbia

Abstract

This study investigated how participants (39 teenagers and 10 college–age online mentors) interacted and constructed shared knowledge of concepts, goals, tasks, procedures, and solutions when solving a real-world problem in an Internet-mediated project-based learning environment (iExpeditions). iExpeditions is a website that provides a coordinated set of communication and collaboration tools, including chat rooms, a discussion forum, "notes" (an intranet email tool), and "logs" (electronic journals). The main focus of this study was on how different patterns of participant interaction contributed to the social construction of shared knowledge. We took a case study approach, examining the knowledge construction process of two of the six online teams involved in the project. The communication artifacts within the two teams were analyzed through psycholinguistic discourse analysis (Edwards & Mercer, 1987). We discovered a three-stage pattern of knowledge construction that occurred in the iExpeditions environment: a) interacting with the expedition guiding materials and online resources, b) interacting with peers online, and c) sharing the knowledge constructed with an outside audience. The findings of this study have implications for the development and implementation of this and other computer supported collaborative learning (CSCL) environments. In particular, our findings point to the importance of the role of the mentor and the opportunity for participants to voice differing ideas as critical to the construction of shared knowledge.

Social constructivist learning theory, with its emphasis on constructing shared knowledge through collaborative work and discourse, has accorded classroom talk a central place in the study of learning (Barnes, 1976; Edwards & Furlong, 1978; Edwards & Mercer, 1987; Edwards & Westgate, 1987; Mehan, 1979). Researchers in the field of computer-supported collaborative learning (CSCL) have called for the study of learner interaction and the development of computer-supported interactive environments to support collaborative learning (Koschmann, 1996; Roschelle & Pea, 1999). This study of the Internet-based Shared Environment for Expeditions (iExpeditions) focused on an analysis of the shared knowledge constructed through online communications among participants and the processes that participants engaged in during the construction. The goal of this study was to discover patterns of interaction mediated through an online environment and their relation to the construction of shared knowledge.

The iExpeditions project is a collaborative research and development project involving the Center for Technology Innovations in Education at the University of Missouri, Brigham Young University, and the Motorola Corporation. The project seeks to develop and understand the use of computer-supported collaborative learning (CSCL) technologies for supporting collaborative problem solving and knowledge construction. The term "expedition" is used as a metaphor for an ill-defined problem-solving activity that takes time to resolve; the term "iExpeditions" refers to the online system used by the participants in completing their tasks. Hereafter we will use the term iExpeditions to refer to the CSCL system and expedition to refer to the participants’ experience in using these tools to work on the telematics problem.

This study was an interpretative case study (Yin, 1989). The iExpeditions project was framed as a single case, with two of the six individual teams as embedded cases. The researcher followed the case study procedure outlined by Yin (1994), which consists of:

• generating the research questions about what the types and contents of shared knowledge are and how shared knowledge is constructed
• identifying the purpose of the study as discovering shared knowledge constructed and the patterns of knowledge construction
• specifying the units of analysis as the iExpeditions community and the individual teams
• establishing the logic linking the data to the propositions
• explaining the criteria for interpreting the findings when writing the results.

The case study method accommodates the research focus of describing and interpreting a phenomenon and meets the twofold research purpose of understanding the phenomenon under study and in generating an explanatory model.

The online communications among the teenage youth and mentors were analyzed using the psycholinguistic discourse analysis method (Edwards & Mercer, 1987). The psycholinguistic approach is based on the sociocultural theory of cognitive development (Vygotsky, 1978). It focuses on the relationship between discourse and the sharing of knowledge (Barnes, 1976; Edwards & Furlong, 1978; Edwards & Mercer, 1987; Mehan, 1979). Although discourse analysis is usually used to study interactions in traditional classroom settings, researchers have also effectively used it in examining online communication (Markel, 1999).

Participants. This six-week expedition was implemented with 39 teenagers between the ages of 13 and 17 who are children of Motorola employees. Forty-five youth were initially selected to participate from those who responded to an invitation. Selection was based on the youths’ self-report of sufficient access to technology to engage in the project. Thirty-nine youth who demonstrated some usage of the system during the first week formed were finally selected to participate in the expedition. Participants met for one three-hour face-to-face orientation, which focused primarily on learning how to use the iExpeditions system.

Ten online mentors, who were students in the teacher education program at Brigham Young University, were recruited to participate in the project. The mentors were recruited based on their interests in the project, their availability during the scheduled time for the expedition, and their previous experiences of mentoring. The mentors went through a one-day training conducted by one of the project stakeholders. They were oriented to the goal and tasks of the expedition, as well as their responsibilities in team building, communicating, problem solving, artifact-generation, and reflecting on team experiences.

Teams. Participants formed six teams of four to five youth and one to two mentors. The team formation was based both on self-selection and the project stakeholders’ random assignment according to their age and gender. With a few exceptions, most of the youth had no prior relationships. With the exception of the orientation, the entire expedition was mediated through the online environment of iExpeditions.

Both of the teams we selected for our case study were composed of two mentors and five teenage participants. The selection was based on two major factors: the difference in the way the two teams were formed and their similarly high level of participation in the expedition. One of the two teams formed primarily through self-selection, because two of the teenage participants were siblings and one other team member was a friend of theirs. The members of the other team had had no contact prior to the expedition. The high level of participation of both teams was evidenced by the number of chat transcripts, notes, logs, and discussion forum messages the teams generated. More than half of these two teams’ members completed individual assignments, participated in the online discussion, and contributed to the progress of the team’s work.

The technological context. iExpeditions was developed to support youth in an extra-curricular experience solving problems situated in a real-life context. An expedition has much in common with project-based learning approaches, such as building learning experiences from problem-driven, collaborative, and extended investigations.

iExpeditions provides users with a coordinated set of communication and collaboration tools, including synchronous tools such as chat rooms and asynchronous tools such as a web-based discussion forum, "notes" (an intranet email tool), and "logs" (electronic journals). iExpeditions was designed to facilitate information access, communication, collaboration, and knowledge representation activities. Using the communications tools that are visible in Figure 1, participants were able to voluntarily exchange email notes, attend online chat sessions, create logs, and post questions, schedules, and any other messages on the forum.

Web-based guiding materials for the expedition were linked to an organizational map that the youth and mentors could easily access within the iExpeditions system. The organizational map outlined the five phases of the problem-solving activities: Phase 1--getting started, Phase 2--gathering data, Phase 3--identifying solutions, Phase 4--developing solutions, and Phase 5--wrapping up. Figure 1 captures the basic interface associated with iExpeditions, including the communication tools on the left side, the members on one’s team on the top, and the problem-solving organizational map at the bottom.

Figure 1. The Basic Interface of iExpeditions.

The Expedition. The CEO of Motorola invited the participants to provide the company with help in developing and marketing telematics. For purposes of this project, "telematics" was defined as wireless communications designed for the car, providing drivers with personalized information, messaging, entertainment, and location-specific travel and security services. Participants worked in teams to develop a proposal to deliver to the CEO and to a panel of telematics experts. The expedition took place over a 6-week time period.

Our observations of participants’ use of the system and collection of communication artifacts were the primary methods for data collection. The communication artifacts were classified into four types: (a) electronic notes exchanged among participants; (b) transcripts of chat room discussions; (c) electronic logs the youth and mentors created in the iExpeditions system, including the team final project presentation logs, and (d) forum entries.

Researcher-designed questionnaires were also used as supplementary data collection instruments for help in understanding youths’ attributions, interest, and participation in the expedition. The questionnaires were distributed by email on a weekly basis and were completed by 85% of the participants in the study. In addition, the researchers conducted semi-structured interviews (see Appendix A) with the youth and mentors on the two sampled teams to follow up on emerging themes. Interviews were conducted through face-to-face communications, chat room discussions, email, and telephone conversations.

We developed a coding scheme (see Appendix B) that included 2 main categories and 12 sub-categories (61 codes) as the primary instrument for discourse analysis. This scheme was used to characterize the flow and structure of the synchronous and asynchronous communications among participants in the expedition. The interrater reliability of two researchers using this coding scheme was 85%. The primary researcher (Wang) and an assistant coded two chat transcripts (10% of all the chat transcripts). We examined the codes assigned to each text unit, counted the frequency of agreement, and then divided the frequency by the total text units (284) coded to get the percentage of agreement (85%). Furthermore, we discussed all the codes not agreed upon and reached consensus on their definitions and functions, to improve the reliability for future coding.

We developed the coding scheme primarily by adapting coding categories from the Adapted Center for Educational Research Innovations (CERI) coding guide, which Harris (1994) has used in analyzing participant interactions in eight classroom-based Motorola expeditions. In addition, we added coding categories from previous research on knowledge construction (Derry, Gance, Gance, & Schlager, 2000; Edwards & Mercer, 1987). The two main categories in the coding scheme encompass codes from the following sources.

Category (1)--Process of Work and Learning was adapted from the Interaction codes in the CERI coding guide. The sub-codes within each category in the iExpeditions coding book were generated from content analysis of this expedition data. Category (1) also includes the codes characterizing the basic elements of a knowledge-construction interaction Edwards and Mercer (1987) identified: (a) offering new information, (b) reference to existing context, (c) requests for information, and (d) checks on the validity of interpretations of information offered. We created two new codes--exchanging resources and information and questioning--to represent these four elements.

Category (2)--Types of Knowledge Construction was derived from the seven types of knowledge construction Derry and her colleagues (2000) hypothesized, based on sociocultural theories (Wertsch, 1991) and the situated cognition perspective (Lave & Wenger, 1991) for assessing online knowledge construction.

The interaction pattern that emerged from our study of participants’ online interactions in iExpeditions includes (a) ways of talking, (b) mentoring strategies, and (c) the result of the interaction.

Ways of talking. In describing the participants ways of talking, we began with the three categories Mercer (1995) outlined in his study of the guided construction of knowledge of schoolteachers and pupils in Open University classrooms and ten English primary schools: disputational talk, cumulative talk, and exploratory talk. Disputational talk is characterized by disagreement and individualized decision-making. In cumulative talk, speakers build positively but uncritically on what the other has said. In exploratory talk, partners engage critically, but constructively with each other, resulting in negotiation and rational debate. These ways of talking are identified by locating discourse features in an interaction (Mercer, 1995; Wang, 2001). For instance, disputational talk is associated with assertions or counter-assertions; cumulative talk is associated with repetition, confirmation, and elaboration; exploratory talk is associated with challenges, counter-challenges, justifications, and consensus reaching. We added one additional way of talking online, individual talk, in which a participant posts a comment, but no one responds to it.

Mentoring strategies. Mentoring strategies were defined in this study as the discourse techniques and other facilitating strategies that mentors used in facilitating individual online activities and teamwork, such as providing support and guidance, mediating negotiation, and reflection. We referred to the three discourse techniques Mercer (1995) described in the classroom setting -- eliciting, responding, and describing -- when analyzing mentoring strategies, and sought to identify the concrete discourse techniques and facilitating strategies mentors used in guiding their teams.

Results of the interaction. Results of the interaction were defined in this study as the eventual outcomes of an interaction instance, which included the shared knowledge constructed and the artifacts the teams produced. Results also included ideas that were not fully developed and artifacts that were started but not completely built.

Through the comparison of the two teams’ knowledge construction activities, including the types of knowledge constructed, when it was constructed, and the frequency of team discussions, we discovered a pattern of knowledge construction in iExpeditions (see Figure 2). We found a framework of knowledge construction similar to that described by Zhu (1998) in a study of online discussions. Zhu’s framework is based on sociocultural constructs and describes a structured sequence of joint knowledge construction. The description of the pattern discovered in this study extends Zhu's framework by providing more detail about activities in each sequence of the knowledge construction pattern.

This pattern explains how the two teams used computer-mediated communications and online activities to create shared team knowledge. The pattern maps to three concentric circles. The inner circle is comprised of the expedition guiding materials and online resources with which participants engage individually. The middle circle represents participants engaged in discussion, thinking, expressing ideas, exchanging views, and making decisions. The major themes of discussion for both teams centered on the weekly problem-solving activities and the expedition guiding materials. The results of the interactions and joint activities -- the shared knowledge constructed, the team artifacts developed or not developed, and the developed and undeveloped ideas -- are represented by the outer circle of the pattern.

The flow of knowledge construction, ideally, is from the inner circle to the outer. Knowledge construction activities began with individual participants work in the inner circle, interacting with the online resources and expedition guiding materials. Individual work in the inner circle included reading background information provided on the web site, creating logs, doing research, and making any other preparations for the online synchronous discussion about the expedition. Expedition guiding materials and any other online resources that represent the background information of the expedition are mapped to the inner circle.

The middle circle represents the multiple zones of proximal development (ZPD) and engagement of the participants. In these ZPDs, participants interacted through various ways of talking, primarily through individual talk, cumulative talk, and exploratory talk. "Individual talk" included notes or forum messages one member posted but no other member respond to. In cumulative talk, team members exchanged ideas and constructively built on each other’s ideas. In exploratory talk, team members negotiated and argued about different points of view, exposed each other to alternative opinions or strategies, and then reached consensus.

The activities of the middle circle usually centered on the themes related to the expedition. We found that the major themes of discussion on both teams were related to the five phases of the expedition activities. The discussions also included themes not related to the expedition work and learning (such as social interactions), which were not analyzed in this study.

The knowledge construction activities eventually led to results, represented in the outer circle -- the shared knowledge constructed, the artifacts the teams collaboratively produced, the artifacts not completed, and the ideas not fully developed. A major artifact of each team was a final PowerPoint presentation delivered online to a Motorola review panel of telematics experts, which was considered one of the major outcomes of the problem-solving activities. Both teams also ended up with abandoned artifacts, such as individual logs that were not built on or were not integrated into team artifacts. For example, one youth on one of the teams created a Team Accomplishment Log in the first week. Because she did not get any feedback from her team members, she deleted the log from her logbook during the third week of the expedition.

Figure 2. The Pattern of Knowledge Construction in iExpeditions. (Note: ZPDS--Zones of Proximal Development)

Themes of Discussion. Most of the themes of discussion centered on the suggested expedition activities. The two teams we focused on shared several themes of discussion (see Table 1). Both teams had discussions about Phase I getting started activities--the iExpeditions tools, the procedures for this expedition, issues pertaining to team formation, and generating the problem statement--as well as about the remaining phases of the expedition--gathering data, identifying and developing solutions, and wrapping up activities.

Table 1. Common Themes of Discussion on Teams A and E

Yet, these two teams varied in the extent they discussed these themes. Because there was also a good deal of variance in the length of notes and chat sessions, we used word count to more accurately reflect the length of these themes of discussion. Table 2 shows the difference in the length of discussion of three sampled themes--expedition procedures, generating the problem statement, and gathering data.

Table 2. Length of Discussion of Three Sampled Themes on Teams A and E

Ways of talking. We sampled 20 instances of interaction from the email notes and chat sessions from each of the two teams and categorized the ways of talking in each instance (Mercer, 1995; Wang, 2001). The interaction instances we selected for analysis were related to the problem-solving tasks. On team A, 10 of the 20 instances were categorized as cumulative talk and seven were exploratory talk. The remaining three instances were categorized as individual talk, messages sent by individuals but not responded to by others. On team E, 17 of the 20 instances were categorized as cumulative talk, one was exploratory talk, and the other two were individual talk.

The predominant way of talking on both teams was cumulative. Team members contributed ideas, listened to others’ ideas, and developed shared team knowledge. The results of the interaction were usually synthesized ideas or decisions.

Although exploratory talk was infrequent on both teams, team A had a higher frequency of exploratory talk, which started with conflicting opinions but ended in consensus. Team A therefore had a greater tendency to explore, negotiate, and debate conflicting ideas, while team E had a greater tendency to uncritically accumulate ideas.

Mentoring strategies. Although the mentors were exposed to some general guides for assuming their various responsibilities, the project stakeholders did not prescribe any particular strategies for them to use in guiding their respective teams. The mentors therefore had leeway to discover and try different strategies. The mentors on the two sampled teams demonstrated different mentoring strategies to guide the teams through the expedition.

The youth on team A did not always engage in or complete the inner circle activities, such as individual reading of the background and guiding materials on the website, the agendas for the chat sessions, or other supportive materials on their own. In many cases, they had not completed their individual preparatory work when they logged on for a chat session. Rather than force them to read the materials in advance, the mentors spent some of the chat time helping them gain the needed information. Then they were able to proceed with the synchronous and asynchronous middle-circle interactions. This mentoring strategy of synchronously helping the youth acquire individual knowledge and strategies used in facilitating team interactions contributed to the teamwork of constructing shared knowledge.

Mentors on team A allowed more social exchanges, without intervening and redirecting talk toward the task at hand. In addition, team A mentors were more frank in challenging the youth and in helping them to voice differing opinions. They took the initiative in challenging or counter-challenging participants’ ideas or suggestions. Their ability to help participants express these differences of opinion led to the high frequency of exploratory talk on team A. Mentoring strategies in mediating the exploratory talk also helped the team resolve the conflicting opinions and reach consensus, essential steps in exploratory talk (Mercer, 1995).

Team E youth were more compliant with the suggestions in the guiding materials to complete their individual work before the synchronous chat sessions. They were better able to use the artifacts of knowledge construction provided for them in the inner circle of the iExpeditions web environment. For instance, during the discussion about generating the problem statement, the mentors sent out six notes and two attachments to the youth to get them to start their individual work in reading the CEO charge. They intentionally encouraged members to share ideas and to collaborate with each other in understanding the reading through posting notes or communication logs. When the team came together for a chat session, the discussion started from the middle circle, in which the youth and mentors exchanged ideas and attempted to construct shared knowledge. Before ending the chat sessions, the mentors also asked the youth to read recaps of the previous chats, milestone materials, the agenda for the current chat, and logs produced by members before returning for the next chat session.

Starting the teamwork from the inner circle helped the youth on team E to acquire some of the declarative knowledge and team shared knowledge and to construct individual knowledge about the project. Having a better grasp of that shared and individual knowledge facilitated middle-circle interactions, resulting in more time for construction of team shared knowledge. The individual activities therefore built foundations for successful knowledge construction as a team. Moreover, the integration of discourse with joint activity in team E created the context for discussion and facilitated the continuity of new knowledge construction.

On the other hand, team E was not as good at negotiation or argumentation skills in exploratory talk. Much like in traditional classroom talk, the youth tended to talk more often to the mentors than to each other, so the mentors had to reflect that information back to the others and establish it as team shared knowledge. This pattern of youth-to-mentor communication and mentor mediation and sharing of information gradually became an interaction norm for team E. In some cases, participants’ spontaneous contributions remained private rather than shared by the team because the mentor did not get the opportunity to share them with the others.

The mentors on team E also appeared to be more oriented toward accomplishing the prescribed team tasks rather than toward facilitating learning through the process of collaborative problem solving. They appeared more polite and encouraging in team chat sessions and rarely challenged the youths' understanding or misunderstandings. Team E also had much more task-oriented communications than socially-oriented ones.

In this study, we investigated knowledge construction of teenagers and college-age mentors on two teams when they were engaged in CSCL-mediated problem-solving activities. We discovered a pattern of knowledge construction in the iExpeditions environment. This pattern explains how the two teams used computer-mediated communications and online activities to create shared team knowledge. The pattern of knowledge construction during the expedition maps to three concentric circles--inner, middle, and outer circles. The three circles act as different lenses for looking at the same phenomena. The inner circle focuses on the individual participants’ interactions with the online resources provided by the iExpeditions environment such as guiding materials and background information. The middle circle focuses on strategies for communication and activities (themes of discussion, ways of talking, and mentoring strategies). The outer circle focuses on outcomes of the communication and activities (artifacts and shared knowledge constructed). Each of these lenses examines participants’ interactions in the social context.

The content of the shared knowledge constructed on the two teams was centered on the expedition activities. The two teams had a good deal of commonality in their themes of discussion. But they differed in the sequence and duration of their knowledge construction.

The observation that team E youth more often completed the inner circle activities of exploring the online resources and expedition guiding materials prior to dialogue with others through chat sessions and posting notes may explain the variance in the length of discussion about the expedition procedures. Participants gained a better understanding of the expedition through those inner circle activities and so needed less time to discuss it. This allowed them more time than their peers on team A to engage in longer discussions in which they worked on generating the problem statement and deciding on ways of gathering data.

The differences highlight the individuality of teamwork and knowledge construction. Although the teams were given a fairly structured process and tools for mediating that process, they appropriated that process and the tools differently in constructing shared knowledge. This finding corroborates the sociocultural notion (Vygotsky, 1978) of the Zone of Proximal Development (ZPD), in which learners engage in a discovery approach to construct shared knowledge. 

The predominance of cumulative talk on both teams corroborates Mercer (1995)’s findings about exploratory talk being sporadic in the primary and middle school classrooms he has investigated. Given the time constraint for developing a presentation of their problem-solving processes and solutions, it may have been easier for the participants to accumulate ideas and get the task accomplished than to engage in prolonged negotiation and debate.

The higher frequency of exploratory talk on team A may have been associated with a stronger social network, build through previous relationships and offline face-to-face interactions Team A was the only team in the expedition that formed through the youths' initiative. Two of the youth were siblings, they were friends with one of the other team members, and they had previously met one of their other two teammates. One of the members with previous connections to three of the other team members took the initiative to get to know the remaining team member and to get him more involved in team discussions. Through this member's efforts, team A youth appeared to build stronger interpersonal and social relations. The youth on this team had more interpersonal interaction. Most chat sessions started with social interactions and the socializing was frequently interwoven into the work discussions. In addition, the social network built on team A appeared to encourage an uninhibited manner of interaction. For instance, when the youth had different opinions, they would say so rather than hold them back. Moreover, although the mentors set the major discussion topics for each chat session, the youth frequently raised other issues spontaneously. Previous research (Bonk & King, 1998; Mercer, 1995) suggests that socially-oriented interaction supports more intimate interpersonal relationships and greater trust in sharing.

The intensive negotiation and argumentation on team A may also reflect the uninhibited and depersonalized effects of communicating in an Internet-based environment, which, as Walther (1996) argued, could facilitate a higher rate of information exchange and strengthen interpersonal and social relations. Several of the youth noted in their self-reports that they felt less constrained and more at ease in voicing different opinions and challenging others’ understandings in online discussions.

Given the challenges to the participants in an expedition and the new and unfamiliar tasks to accomplish in a short time with a new set of tools, online mentors are a very important resource. The mentors in this expedition did a good job in guiding and supporting the participants, but some strategies seemed better in helping them accomplish their goals than others. For instance, the mentors succeeded in getting the youth to send email notes and participate in chat sessions, but did not succeed in getting them to create more team artifacts, such as logs, to capture the shared knowledge constructed. In addition, the mentors did not guide the youth to use the archived background information and resources provided by the telematics experts, which would have, no doubt, improved the problem-solving process and the problem solutions.

Moreover, we found certain mentoring strategies seemed to optimize knowledge construction among the youth. The mentors on both teams also appeared to encourage and reward cumulative talk, frequently giving positive feedback and uncritically accepting participants’ ideas. The mentors' accepting and positive attitudes, no doubt, contributed to the predominance of cumulative talk.

Previous research (Derry & DuRussel, 1999; Derry, Gance, Gance, & Schlager, 2000) suggests that exploratory talk has a higher value for knowledge construction. The higher frequency of exploratory talk on team A may be partially attributed to the mentors’ efforts in challenging youth and in eliciting different ideas. The lack of exploratory talk on team E may have been related to the mentors’ uncritical acceptance of the youths’ contributions. While this strategy may have been used as a way to avoid discouraging the youth from giving more input, we recommend that mentors should take the initiative in starting negotiations or involving team members in reasoned argumentation. Exploration is encouraged when the mentors value and model clarifying confusion and reconciling conflicts during shared knowledge construction.

This study investigated how participants in a CSCL-mediated project-based learning environment interacted and constructed shared knowledge of concepts, goals, tasks, procedures, and solutions when solving a real-world problem. The main focus was on how participant interaction contributed to the social construction of shared knowledge. The findings of this study have implications for the development and implementation of CSCL environments with participants of similar age groups. In particular, our findings point to the important role of the mentor in encouraging participants to voice their own ideas and negotiate meaning and consensus as a part of the process of knowledge construction. While there is still much to be learned about exploratory talk and the strategies and CSCL tools that best encourage it, the current study adds to a growing set of reports that argue for increasing the amount of exploratory talk in collaborative learning.

The vision of Motorola University in developing Expeditions was primarily to enable young people to work with tools and people at a distance to produce new ideas and products in a way that would encourage participants to be active problem-solvers who are able to work on teams. The Internet-based Expeditions has the potential to support active learning on-line involving Motorola professionals and youth who are geographically dispersed, both within the US and throughout the world. In addition, we believe that our findings will also be valuable for other organizations that are actively seeking ways to engage their members in problem-solving tasks through Internet communications.

How teams interacted in the middle circle and how mentors facilitated the interactions were critical factors in shaping the quality of shared knowledge construction. The findings of this study suggest that the ways of talking used by the participants are related to the quality of shared knowledge constructed. Moreover, mentoring strategies used to facilitate team interactions are related to ways of talking of participants. Further analysis is needed with the current data to examine the relationships among the results of teamwork and the various ways of talking and mentoring strategies of participants. Further studies should also examine team formation strategies and the presence of social talk and their relation to building stronger ties and greater trust among team members.

The authors wish to thank Dr. Paul Wangemann from Motorola University for his support of this study. His vision in developing the Internet-based Expeditions created the opportunity for this study and his generosity in supporting the primary researcher in conducting the fieldwork helped to make this study a success.

Barnes, D. (1976). Language in the secondary classroom. In D. Barnes, J. Britton, & H. Rosen (Eds.), Language, the learner and the school. Harmondsworth: Penguin.

Bonk, C. J. and King, K. S. (Eds.), Electronic collaborators: learner-centered technologies for literacy, apprenticeship, and discourse. Mahwah, NJ: L. Erlbuam Associates.

Derry, S. J., and DuRussel, L. A. (July 1999). Assessing knowledge construction processes in on-line learning communities. Paper presented at the AIED99, Conference of the International Society for Artificial Intelligence, Lemans France.

Derry, S. J., Gance, S., Gance, L. L., and Schlager, M. (2000). Toward assessment of knowledge building practices in technology-mediated work group interactions. In S. P.Lajoie (Ed.), No more walls: Theory change, paradigm shifts and their Influence on the use of computers for instructional purposes (Vol. II). Mahwah, NJ: Erlbaum.

Edwards, A. D., and Furlong, V. J. (1978). The language of teaching. London: Heinemann.

Edwards, D., and Mercer, N. (1987). Common knowledge: The development of understanding in the classroom. London: Methuen.

Edwards, A. D., and Westgate, D. P. G. (1987). Investigating classroom talk (2nd ed.). Washington, D.C.: Falmer Press.

Harris, R. C. (1994). CERI/Motorola active learning: A case study report (Motorola Internal Report). Provo: Brigham Young University.

Koschmann, T. (Ed.). (1996). CSCL: Theory and practice of an emerging paradigm. Mahwah: Lawrence Erlbaum Associates.

Lave, J., & Wenger, E. (1991). Situated Learning: Legitimate peripheral participation. Cambridge: Cambridge University Press.

Markel, S. (1999). Gendered voices: Provocateur in an on-line virtual conference course for in-service teachers. Paper presented at the ED-MEDIA 99, Seattle, Washington.

Mehan, H. (1979). Learning lessons: Social organization in the classroom. Cambridge, MA: Harvard University Press.

Mercer, N. (1995). The guided construction of knowledge: Talk amongst teachers and learners. Philadelphia: Multilingual Matters LTD.

Roschelle, J., and Pea, R. (June-July 1999). Trajectories from today's WWW to a powerful educational infrastructure. Educational Researcher, 28(5), 22-25.

Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes (M. Cole, V. John-Steiner, S. Scribner, & E. Souberman, Eds. and Trans.). Cambridge, MA: Harvard University Press.

Walther, J. (1996). Computer-mediated communication: Impersonal, interpersonal, and hyperpersonal interaction. Communication Research, 23(1), 3-43.

Wang, M. J. (2001). The Construction of Shared Knowledge in An Internet-Based Shared Environment for Expeditions (iExpeditions): a Study of External Factors Implying Knowledge Construction. Unpublished doctoral dissertation, University of Missouri, Columbia.

Wertsch, J. V. (1991). Voices of the mind: A sociocultural approach to mediated action. Cambridge, Mass. : Harvard University Press.

Yin, R. (1994). Case study research: Design and methods. (2nd ed.). Thousand Oaks: Sage

Zhu, E. (1998). Learning and mentoring: Electronic discussion in a distance-learning course. In C. J. Bonk and K. S. King (Eds.), Electronic collaborators: learner-centered technologies for literacy, apprenticeship, and discourse . Mahwah, NJ: L. Erlbuam Associates.

Step 1 Review with the mentors the steps that the team has gone through in making the final presentation.

1. Describe how your team made the final presentation: When did you start to work on the final presentation? How was the task divided? How much did each youth contribute? How long did it take?
2. When did you start to introduce your team to the presentation sample and How is your team product influenced by the presentation samples?
3. What are the major feedback you got from the experts and what did you team learn from this chat?
4. About procedure:

• Did your team really understand what they were doing, or simply wanted to get things done?
• How were the team decisions made?

5. Were there any disagreements? How did your team reach consensus?

Step 2 Review with mentor their chat sessions.

1. Were there lots of knowledge-sharing talks?
2. Did you team build shared knowledge as a result of the talk?
3. And how did you know when your team build shared knowledge?
4. At what point of the Expedition did you see your team actively shared their ideas and really made sense of the teamwork?
5. How frequent has this happened?
6. When team members came to chat unprepared, how did you get them started with the teamwork?
7. Did you often recall what your team has understood before when doing a new activity?

Coding Tree

Category (1) Process of Work & Learning

(1 1) Y interact(s) M (+)
      (1 1 2)exchanging resources & information
      (1 1 4)individual contributions
      (1 1 5)disagreements & confusion(*)
      (1 1 6)questionning (challenging)(*)
      (1 1 7)explain ideas(*)
      (1 1 8)reach consensus(*)
      (1 1 10)suggesting alternatives(*)
      (1 1 11)social interactions

(1 2) Y interact(s) Y (+)
      (1 2 2)exchanging resources & information
      (1 2 5)disagreement & confusion(*)
      (1 2 6)questionning (challenging)(*)
      (1 2 7)explain ideas(*)
      (1 2 8)reach consensus(*)
      (1 2 9)getting organized

(1 3) M interacts Y
      (1 3 1)giving or receiving help
      (1 3 2)exchanging resources & information
      (1 3 5)disagreements & confusion(*)
      (1 3 6)questionning(challenging*)
      (1 3 7)explain ideas(*)
      (1 3 8)reach consensus(*)
      (1 3 9)team building
      (1 3 10)problem solving
      (1 3 11)artifacts generating

(1 4) M interacts M
      (1 1 1)giving(receiving) help
      (1 1 2)social interactions
      (1 1 3)learning to be mentor
      (1 1 4)exchanging resources & information
      (1 1 6)questioning

(1 5) individual work (+)
      (1 5 1)none or incomplete
      (1 5 2)complete
      (1 5 3)work on challenges
      (1 5 4)search information
      (1 5 5)check own answers

Category (2) Types of knowledge construction

    (2 1) Type 1: eliciting or giving input/feedback   

        (2 1 1) Elicit
            (2 1 1 1) direct elicitation
            (2 1 1 2) cued elicitation

        (2 1 2) Respond
            (2 1 2 1) confirmations
            (2 1 2 2) rejections
            (2 1 2 3) repetitions
            (2 1 2 4) reformulations

        (2 1 3) Describe
            (2 1 3 1) ‘we’ statement
            (2 1 3 2) literal recaps
            (2 1 3 3) reconstructive recaps

    (2 2) Type 2: synthesizing ideas

    (2 3) Type 3: non-verbal behaviors
        (2 3 1) reading (notes, chat agenda etc.)
        (2 3 2) opening browser windows
        (2 3 3) sending log, notes
        (2 3 4 ) searching information
        (2 3 5) all other activities

    (2 4) Type 4: joint product development
        (2 4 1) generating artifacts
        (2 4 2) assembling individual logs

    (2 5) Type 5: apprenticeship   
        (2 5 1) explanations
        (2 5 2) repetitions
        (2 5 3) confirmations

    (2 6) Type 6: negotiation
        (2 6 1) disagreement
        (2 6 2) questioning
        (2 6 3) challenging
        (2 6 4) consensus

    (2 7) Type 7: exchanging resources and information