From Jsarmi

Facilitating and analysing roles in computer-supported collaborative learning (EARLI 2007 symposium)

Organisers Jan-Willem Strijbos (Leiden University, The Netherlands), Armin Weinberger (Ludwig Maximillians University, Germany Discussant Sanna Järvelä (Oulu University, Finland)

Summary

During the past five years, the ‘role’ concept has become a promising construct for facilitating and
analysing Computer-Supported Collaborative Learning (CSCL). Roles can be defined as more or less
explicit functions or responsibilities to guide individual members’ behaviour and regulate their group
interaction (Hare, 1994). Roles can promote awareness of the overall group performance and member’s
contribution (Strijbos, Martens, Jochems, & Broers, 2004). However, the role distribution is not always
optimal from a learning viewpoint, and some students are more inclined to take up a role voluntarily. In
CSCL research roughly two perspectives regarding roles exist: scripted roles that aim to facilitate
collaborative learning processes and roles that emerge spontaneously developed by the participants in
support of their collaborative learning activities.

Scripted roles perspective. Roles can be assigned to learners as instructional support to structure the
collaborative process; this is also referred to as ‘scripting’ (O’Donnell & Dansereau, 1992; Dillenbourg
2002; Weinberger, 2003). These roles are either content-oriented or process-oriented. A content-oriented
role, like a summarizer, has been found to stimulate higher levels of knowledge acquisition (Schellens,
Van Keer, & Valcke, in press). A process-oriented role specifies members’ individual responsibilities to
stimulate coordination (Strijbos et al., 2004).

Emergent spontaneous roles perspective. Roles can emerge spontaneously without scripting being
provided (Strijbos, De Laat, Martens, & Jochems, 2005), and the concept of roles can be applied to
analyze the individual contributions and interaction patterns during collaborative learning (De Laat, 2005;
Hermann, Jahnke, & Loser, 2004; Pilkington & Walker, 2003).

This symposium will focus on both perspectives – the scripting of roles for instructional support and the
analysis of spontaneous roles during collaborative learning. The symposium contributions will address the
feasibility of these applications and explore how these seemingly diverse perspectives can be integrated.

Prototypical roles in group work: A conceptual framework for the design of a tool for teachers Jan-Willem Strijbos Leiden University The Netherlands jwstrijbos@fsw.leidenuniv.nl Maarten De Laat University of Exeter United Kingdom m.f.delaat@exeter.ac.uk

Abstract During the past five years, the ‘role’ concept has become a promising construct for facilitating and analysing Computer-Supported Collaborative Learning (CSCL). In CSCL research there are roughly two perspectives regarding roles: scripted roles that aim to facilitate collaborative learning processes and emergent roles developed spontaneously by the participants in support of their collaborative learning activities.

Scripted roles can be assigned to learners to structure the collaborative process; this is also referred to as ‘scripting’ (O’Donnell & Dansereau, 1992), and are content-oriented or process-oriented. A contentoriented role, e.g. summarizer, appears to stimulate knowledge acquisition (Schellens, Van Keer, & Valcke, 2005), whereas a process-oriented role specifies members’ individual responsibilities to stimulate coordination (Strijbos, Martens, Jochems, & Broers, 2007). Roles can also emerge spontaneously without any scripting being provided (De Laat, 2006). The concept of roles can then be applied to analyze students’ individual contributions and interaction patterns during collaborative learning (Pilkington & Walker, 2003).

The recent methodological development and discussion signifies that we as researchers struggle to find satisfactory ways to make sense of these complex interactions. However this is not an end in itself. Our efforts should not only serve our analytical and theoretical needs, but should also be transformed into ‘tools for teachers’ to provide them with the adequate support needed to implement CSCL in the classroom.

In this contribution a conceptual framework consisting of prototypical roles, defined along three dimensions, is presented to describe student activity in both small groups (where roles are frequently scripted) and large groups (where roles are frequently spontaneous). The goal is to provide a framework that a) facilitates a meaningful description of student behaviour, and b) can be handled with relative ease by teachers to evaluate group work.

Extended summary During the past five years, the ‘role’ concept has become a promising construct for facilitating and analysing Computer-Supported Collaborative Learning (CSCL). Roles can be defined as more or less explicit functions or responsibilities to guide individual members’ behaviour and regulate their group interaction (Hare, 1994).

In CSCL research roughly two perspectives regarding roles exist. The scripted roles that aim to facilitate collaborative learning processes, also referred to as ‘scripting’ (O’Donnell & Dansereau, 1992), and the emergent roles (developed by participants during collaborative work) that are applied to analyze collaborative learning processes (De Laat, 2006; Pilkington & Walker, 2003). In addition to the rise of the role concept to facilitate and analyse the student behaviour in CSCL, it is also apparent that there is a lack of tools for teachers. The recent methodological development and discussion signifies that we as researchers struggle to find satisfactory ways to make sense of these complex interactions. Our efforts should not only serve our analytical needs, but the large teacher community as well. Hence, transforming analysis techniques into ‘tools for teachers’ is the logical next step. In this contribution a conceptual framework consisting of prototypical roles to a) facilitate a meaningful description of student behaviour during group work, and b) can be handled with relative ease by teachers to moderate and evaluate group work. Prototypical roles

Our conceptual framework consists of eight prototypical roles, which are defined along three dimensions (see Figure 1). The first dimension consists of the group size: behaviour in a small group (three to six members) is different from that in a large group (seven or more) (Forsyth, 1999). The second dimension consists of the students’ orientation during group work: either a student is oriented towards individual goals or towards group goals. Hakkarainen, Lipponen, Järvelä and Niemivirta (1999) distinguish three learning orientation dimensions of which two are particularly relevant with respect to the proposed framework: the ‘autonomy’ dimension “stresses students’ opportunities to take leadership roles, develop a sense of personal control and autonomy in the learning process” (p. 265), and the ‘grouping dimension’ which “focuses on the students ability to work effectively with others on school tasks” (p. 265). The third and final dimension consists of the effort that students invest in the group work. However, effort is not the same as ‘impact’, meaning that even group member with few contributions can still be very influential in the group. Nevertheless, effort is relatively easier to determine than impact. Considering these three dimension, eight prototypical roles can be defined (Figure 1).

Figure 1. Prototypical roles.

The top half shows prototypical roles for a small group and the bottom half for a large group. On the right side are the group oriented roles, and roles with individual orientation are on the left side. Each quadrant is divided in low versus a high effort. The horizontal axis is a mirror for the comparison of small and large groups, for example the ‘Captain’ in a small group has the ‘Pilar’ in a large groups as its mirror image. This conceptual framework also contains two concepts that are frequently used to typify students’ behaviour: the ‘Free-rider’ and ‘Lurker’. In addition, the framework reveals typical behaviour that has not been specifically addressed thus far in the literature: the ‘Over-rider’ and ‘Pusher’. Students occupying these roles strive to realise their personal goals rather than the group’s goals and will invest considerable effort to direct the group orientation towards their personal goals (or make them a prominent part of the group goals).

Qualitative analysis A detailed exploratory study by De Laat (2006) on emergent student roles in CSCL revealed the complexity of participation in group learning activities. First of all the study highlighted the impact of the task on how the students structure their collaborative activities, and 2) that students develop a personal style (emergent roles) during online learning activities. Second, this study revealed how students develop a social meta-cognitive awareness of each other’s learner styles and use this to structure and regulate group learning.

Several examples of the described prototypical role behaviour will be drawn from two distinct datasets, one where roles were scripted in a small group (Strijbos et al., 2007) and one where roles emerged spontaneously in a large group (De Laat, 2006), to illustrate the viability of the prototypical role conceptual framework.

Conclusion and discussion The analysis of the prototypical roles will advance our understanding of group work (CSCL) from a descriptive/bottom up approach. The analysis of prototypical roles will support both development of scripted roles and analysis of emergent roles in future collaborative learning scenarios. More importantly, it provides a framework for meaningful description of students’ behaviour, to enable teachers to evaluate group work with relative ease.

References De Laat, M. F. (2006). Networked learning. Unpublished doctoral dissertation, Utrecht University, Utrecht, The Netherlands. Forsyth, D. R. (1999). Group dynamics (3rd ed.). Belmont: Wadsworth. Hakkarainen, K., Lipponen, L., Järvelä, S., & Niemivirta, M. (1999). The interaction or motivational orientation and knowledge seeking inquiry in computer-supported collaborative learning. Journal of Educational Computing Research, 21, 263-281. Hare, A. P. (1994). Types of roles in small groups: A bit of history and a current perspective. Small Group Research, 25, 443-448. O’Donnell, A. M., & Dansereau, D. F. (1992). Scripted cooperation in student dyads: A method for analysing and enhancing academic learning and performance. In R. Hertz-Lazarowitz & N. Miller (Eds.), Interaction in cooperative groups: The theoretical anatomy of group learning (pp. 120-144). New York: Cambridge University Press. Pilkington, R. M., & Walker, S. A. (2003). Facilitating debate in networked learning: Reflecting on online synchronous discussion in higher education. Instructional Science, 31, 41-63. Schellens, T., Van Keer, H., & Valcke, M. (2005). The impact of role assignment on knowledge construction in asynchronous discussion groups: A multilevel analysis. Small Group Research, 36, 704- 745. Strijbos, J. W., Martens, R. L., Jochems, W. M. G., & Broers, N. J. (2007). The effect of functional roles on perceived group efficiency during computer-supported collaborative learning: A matter of triangulation. Computers in Human Behavior, 23, 353-380.

Role scripts for improving group learning beyond individual learning: Does it work? Armin Weinberger Ludwig Maximillians University Germany armin.weinberger@psy.lmu.de Karsten Stegmann Ludwig Maximillians University Germany karsten.stegmann@psy.lmu.de Frank Fischer Ludwig Maximillians University Germany frank.fischer@psy.lmu.de

Abstract Studies show that computer-supported collaborative learners frequently suffer from process losses and seem to have difficulties to distribute roles effectively. Hence, individual learning may outperform collaborative learning if learners are not supported to take over complementary roles. Computer-supported scripts seem to be able to facilitate specific processes and outcomes of computer-supported collaborative learning (CSCL), such as domain-specific knowledge and domain-general, e.g., argumentative knowledge. In this study with a 2×2-factorial design (n = 72) we investigate the effects of a script (with vs. without) and the social form of learning (individual vs. collaborative) on learning processes and outcomes in the context of an online learning environment in higher education. Results show that the script facilitated the construction of arguments within the computer-supported learning environment as well as knowledge acquisition. The results indicate that collaborative learning may outperform individual learning regarding learning outcomes when it is structured by a script.

Extended summary Learning partners are regarded as learning resources, e.g., by providing additional knowledge to solve the learning task and by taking over specific functional roles and learning activities. For instance, a learner may take over the role of a case analyst and contribute a provisional problem analysis. Other learners may take over the role of critics and modify the initial analysis. Structuring computer-supported collaborative learning (CSCL) according to this recurrent analyst-critic-pattern facilitated learning outcomes beyond what could be achieved with unstructured CSCL (Weinberger, Ertl, Fischer, & Mandl, 2005). Apparently, collaborative learning can be structured in a way that important learning processes are being effectively distributed over a group of learners. Hence, the guided distribution of specific roles – as it is being realized in scripted cooperation (O’Donnell, 1999) – seems to facilitate collaborative learning. What makes scripts successful instructional tools for collaborative learning? Do scripts, for instance, reduce process losses typically experienced in CSCL, such as coordination problems (Strijbos, Martens, Jochems, & Broers, 2004), or do they rather support meaningful learning activities of the individual learner, such as constructing sound arguments (Stegmann, Wecker, Weinberger, Fischer, 2006)? In the later case, scripts may facilitate individual learning to a similar extent as collaborative learning. RQ1 is to what extent does a script (with vs. without) and the social form of learning (individual vs. collaborative) and the combination thereof affect processes of computer-supported learning? We hypothesize that the script would foster the processes of individual and collaborative computer-supported learning.

To devise complex instructional arrangements for CSCL does not seem to be worth the effort compared to individual learners who do not suffer from specific process losses of CSCL and who may be better acquainted with individual learning scenarios. It has been argued, though, that the social form of collaborative learning can outperform individual learning in fostering not only domain-specific knowledge, but also domain-general knowledge, e.g., argumentative knowledge (Stegmann et al., 2006). RQ2 is to what extent does a script (with vs. without) and the social form of learning (individual vs. collaborative) and the combination thereof affect the individual learning outcomes? We hypothesize that the script would foster learning outcomes of collaborative learners beyond the level that unscripted collaborative and individual learners would attain.

Method In this experimental study with a 2×2-factorial design (n = 72) we investigate the effects of a script (with vs. without) and the social form of learning (individual vs. collaborative) on learning processes and outcomes in the context of a computer-supported learning environment in higher education. Learners were to analyze problem cases with the help of an attribution theory individually or in groups of three. The script aimed to support the role of an argument constructor by suggesting learners to explicate claims, data, warrants, and qualifiers. The script was implemented in the interface of the computer-supported learning environment and provided text windows for each argument component. Learning processes were analyzed with respect to the formal quality of arguments (i.e. the frequency of warranted and qualified claims) and the epistemic quality of arguments within the learning environment (i.e. the frequency of arguments that contributed to solving the learning task by applying specific theoretical concepts adequately to a problem case). Learning outcomes were analyzed with respect to domain-specific knowledge (i.e. the extent to which learners were individually able to apply specific theoretical concepts to a transfer problem case after participating in the computer-supported learning session) and argumentative knowledge (i.e. the extent to which learners were individually able to recall argument components such as claim, warrant, and qualifier and to construct warranted and qualified claims on another topic in a pen-andpaper post-test). All process and outcome variables had been measured reliably.

Results Regarding RQ1, the findings show clearly that the script reduces epistemic, but facilitates formal quality of arguments. Regarding RQ2, formerly scripted collaborative learners acquired more domain-specific and argumentative knowledge than any other experimental group (see table 1). Table 1. Learning outcomes in z-scores of domain-specific and argumentative knowledge for each experimental group. Individual learning Collaborative learning Without script With script Without script With script Domain specific knowledge M 0.19 -.051 -0.29 0.60 SD 1.02 0.64 0.71 1.25 Argumentative knowledge M -0.80 -0.51 -0.51 0.94 SD 0.77 0.64 0.64 0.63

Discussion The findings indicate that scripts can help both computer-supported individual and collaborative learners to engage in specific meaningful learning activities, but that particularly computer-supported collaborative learners benefit from scripts – possibly by additionally reducing process losses of learning together in CSCL environments. These results give rise to the following assumptions: Unscripted collaborative learners do not seem to be able to distribute roles in a complementary and effective way themselves. Collaborative learners may lack knowledge on how to overcome barriers and learn effectively together (Strijbos et al., 2004). CSCL unfolds its potential only, when the degree of freedom is not too large (Kirschner, Sweller, & Clark, 2006). Distributing roles by scripting learners seems to be a feasible way to particularly enhance collaborative learning and harvesting the promised added value of learning together.

References Kirschner, P. A., Sweller, J., & Clark, R. E. (2006). Why minimal guidance during instruction does not work: An analysis of the failure of constructivist, discovery, problem-based, experiential, and inquiry-based teaching. Educational Psychologist, 41(2), 75-86. O'Donnell, A. M. (1999). Structuring dyadic interaction through scripted cooperation. In A. M. O'Donnell & A. King (Eds.), Cognitive perspectives on peer learning (pp. 179-196). Mahwah, NJ: Erlbaum. Stegmann, K., Wecker, C., Weinberger, A., & Fischer, F. (2006). Collaborative argumentation and cognitive processing: An empirical study in a computer-supported collaborative learning environment. Strijbos, J. W., Martens, R. L., Jochems, W. M. G., & Broers, N. J. (2004). The effects of functional roles on group efficiency: Using multilevel modeling and content analysis to investigate computer-supported collaboration in small groups. Small Group Research, 35, 195-229. Weinberger, A., Ertl, B., Fischer, F., & Mandl, H. (2005). Epistemic and social scripts in computersupported collaborative learning. Instructional Science, 33, 1-30.

Roles as structuring tool in online discussion groups: Studying students’ role performance and the impact on knowledge construction Bram De Wever Gent University Belgium bram.dewever@ugent.be Tammy Schellens Gent University Belgium tammy.schellens@ugent.be Hilde Van Keer Gent University Belgium hilde.vankeer@ugent.be Martin Valcke Gent University Belgium martin.valcke@ugent.be

Abstract This contribution connects the results of three successive studies on the introduction of roles as a scripting tool in order to structure asynchronous discussion groups. The studies fit in with the search for instructional approaches stimulating social knowledge construction in e-discussions. More specifically, this contribution fits in with the abovementioned scripted roles perspective. The introduction of five roles was examined: starter, summariser, moderator, theoretician, and source searcher.

The studies were conducted in a naturalistic higher education setting with freshmen enrolled in the course Instructional Sciences. Asynchronous discussion groups of ten students were organised to foster students’ processing of the learning content. Four successive authentic tasks of three weeks each were presented and roles were assigned to students in order to promote knowledge construction through social negotiation.

This contribution focuses first on exploring whether students enacted the roles they were assigned. Subsequently, the impact of assigning roles on knowledge construction is studied. In order to determine the level of knowledge construction, content analysis based on the interaction analysis model of Gunawardena, Lowe, and Anderson (1997) was performed. Taking into account the hierarchical nesting of students in discussion groups and the successive nature of the themes, repeatedmeasures multilevel modelling was applied.

Concerning role assignment, the results confirm that all students enact the assigned roles. Moreover, although source searchers, theoreticians, summarisers, and students without roles in the role condition focused to a lesser extent on some activities related to other roles, students generally did not neglect other activities.

Concerning the impact of roles, this contribution shows a positive effect on students’ social knowledge construction, especially when roles are used as a structuring tool at the beginning. This implies that the moment of introduction is important: roles should be introduced at the start of the discussions and can be faded out towards the end.

Extended summary This study fits in with the search for instructional approaches to stimulate knowledge construction through social negotiation in online asynchronous discussions. The study is situated in the context of a first year course Instructional Sciences, where asynchronous discussion groups of 10 students are organised to foster students’ processing of the learning content. In order to promote knowledge construction through social negotiation, roles are assigned to students. The main goal of this study is to explore the impact of the introduction of roles on students’ level of knowledge construction in asynchronous discussion groups. Roles can be considered a specific type of scripting that is used to structure and improve collaborative discourse. In the present study the performance of 5 different process-based roles is studied: starter, summariser, moderator, theoretician, and source searcher.

The first research question explores whether students enacted the roles they were assigned. The second research question studies the effect of role support. More specifically, the impact of the moment of introduction of the role support on students’ knowledge construction through social negotiation is examined. The third research question studies the differential impact of the different roles on the level of social knowledge construction reflected in students’ messages.

Method For the first research question, focusing on analysing the role-related activities in students’ contributions, an analysis model was developed to identify message characteristics on five different dimensions: sources, theory, summaries, moderation, and new points. These dimensions are related to the assigned roles. Ordinal, multinomial, and logistic regression procedures were executed to study whether students enacted their roles.

For the second research question, two conditions were contrasted. In the first condition role support was only implemented during theme 3 and 4. In the second condition, roles were implemented during theme 1 and 2. Roles were randomly assigned and rotated after the first role-supported discussion, implying that each student was assigned a role once.

For the second and third research question, content analysis based on the model of Gunawardena, Lowe, and Anderson (1997) was performed in order to determine the level of knowledge construction reflected in students’ messages. Taking into account the hierarchical nesting of students in discussion groups and the successive nature of the themes, repeated-measures multilevel modelling (Hox, 1998, Snijders & Bosker, 1999) was applied to study the impact of the moment of the role introduction and the differential impact of the different roles.

Results Concerning the first research question, the results clearly indicate that students perform the roles they were assigned (see the plusses on the main diagonal in Table 1). Although they pay less attention to some activities (see the minuses in Table 1), they generally do not neglect the activities related to the other roles while discussing (see the equals signs in Table 1).

Table 1. Overview of the results concerning role performance Role Source Theory Summary Moderation New points Source searcher +++ = = = / = / n.c. = Theoretician = +++ = = / = / n.c. = Summariser – = +++ = / = / n.c. – – – Moderator = = = ++ / +++/ + = Starter = = = = / + / ++ +++ No role = = – – – / = / = = No role condition Reference category Reference category Reference category Reference category Reference category Moderating column: Organisational / content / both forms of moderating / n.c.: not calculated +++ and – – – < .001/6 = .00017, ++ and – – < .01/6 = .0017, + and – < .05/6 = 0.0083 With regard to the second research question, the results point to a significant difference in levels of social knowledge construction in theme 1 and 3. In both themes, contributions in discussion groups in which roles were introduced at the start (condition 2, role/no-role) reflect higher levels of knowledge construction compared to discussion groups in which roles are introduced halfway (condition 1, no-role/role). The results are depicted in Figure 1. Figure 1. Impact of conditions on social knowledge construction. Table 2. Overview of the differential impact of the roles on the levels of social knowledge construction (LKC) Role LCK Source searcher = Theoretician ++ Summariser +++ Moderator ++ Starter = No role + No role condition Reference category +++ < .001, ++ < .01, + < .05

Concerning the third research question, Table 2 presents an overview of the differential impact of the different roles on the level of knowledge construction, indicating that contributions of theoreticians, summarisers, moderators, and students without roles in a discussion group with roles reflect higher levels of knowledge construction. No negative effects were found.

Discussion and conclusion First, we can conclude that the scripting approach presented is fruitful. All students perform the assigned roles, implicating that assigning roles is a recommended scripting approach to structure asynchronous discussion groups and stimulate students to perform valuable activities. Moreover, roles can be an important structuring tool to enhance students’ level of social knowledge construction if they are introduced at the start of the discussions and faded out afterwards. In this respect, it can be noticed that groups with role support at the beginning outperform the others in both theme 1 and 3, even when role support was cut back (theme 3). Last, assigning roles to students has no negative impact on the levels of knowledge construction reflected in their contributions. On the contrary, contributions of theoreticians, summarisers, moderators, and students without roles in a discussion group with roles reflect higher levels of knowledge construction.

References Gunawardena, C. N., Lowe, C. A., & Anderson, T. (1997). Analysis of a global online debate and the development of an interaction analysis model for examining social construction of knowledge in computer conferencing. Journal of Educational Computing Research, 17, 397-431. Hox, J. J. (1998). Multilevel modeling: When and why. In R. Balderjahn, R. Mathar & M. Schader (Eds.), Classification, data analysis, and data highways (pp. 147-154). New York: Springer-Verlag. Snijders, T. A. B., & Bosker, R. J. (1999). Multilevel analysis. London: Sage.

Boundaries and roles: Social location and bridging work in the Virtual Math Teams (VMT) online community Johann Sarmiento Drexel University United States jsarmi@drexel.edu Wesley Shumar Drexel University United States shumarw@drexel.edu

Abstract As research in Computer-Supported Collaborative Learning (CSCL) expands its understanding of diverse types of joint learning activities and the participation frameworks they enact, new perspectives on how social reality is constructed become necessary for analysis and design purposes. Our research concentrates on the temporal development of online learning groups into teams and communities and the interactional emergence of positioning or situated roles —dynamic orientations toward participation in small-group interaction. We investigate the ways that small virtual teams engaged in sustained work over time cross over the boundaries of time, episodes, collectivities, and perspectives to constitute and advance learning and knowledge-building as a continuous activity. We refer to this interactional activity as "bridging" work. While engaged in bridging work, team members position themselves, their co-participants and other collectivities dynamically in ways that suggest the need to consider the “situatedness” aspect of the concept of role in CSCL research. Bridging activity, and the positioning work it entails, contributes to the construction and maintenance of a joint problem space over time, to manage ongoing participation, and to the constitution of the temporal imagination of the collectivities involved.

Extended summary As research in CSCL expands its understanding of diverse types of joint learning activities and the participation frameworks they enact, new perspectives on how social reality is constructed become necessary. Our research concentrates on the temporal development of online learning groups into teams and the interactional emergence of situated roles —dynamic orientations toward participation in smallgroup interaction. In fact, the concept of role —as an attribute of an individual— although instrumental in small-group research (e.g., Hare, 2003) and in CSCL (e.g., Strijbos et al., 2007), has been questioned recently in other areas of research in favor of a more situated perspective (e.g., Davies & Harre, 1990). This perspective favors the actual discursive processes of “positioning” and locating social participants in conversations and interactions. We approach our analysis from this interactional perspective in order to explore the situated perspective of roles within the temporal development of learning teams in the context of the Virtual Math Teams (VMT) online community.

The VMT project at the Math Forum investigates the innovative use of online collaborative environments for mathematics learning (Stahl, 2005). The Math Forum is an online math community, active since 1992. It promotes technology-mediated interactions among teachers, students, mathematicians, staff members and others interested in learning, teaching, and doing mathematics. Central to the VMT research program are the investigation of the nature and dynamics of group cognition (Stahl, 2006) as well as the design of effective technological supports for small-group interactions. In this particular study we investigate the ways that small virtual teams engaged in sustained work over time, crossing over the boundaries of episodes, collectivities, and perspectives to constitute and advance learning and knowledge-building as a continuous activity. We refer to this interactional activity as "bridging" work. Bridging is achieved through a set of methods through which participants deal with the discontinuities relevant to their collective engagement. Bridging thereby might tie events at the local smallgroup unit of analysis to interactions at larger units of analysis (e.g. online communities, multi-team collectivities, etc.) as well as between the individual and small-group levels.

Data sources and analysis During the spring of 2005, an experiment was conducted to explore collaborative knowledge building over time. Five virtual teams were formed with about four non-collocated secondary students selected by volunteer teachers at different schools across the USA. The teams engaged in online math discussions for four hour-long sessions over a two-week period. In the first session, teams were given a brief description of a non-traditional geometry environment: a grid-world where one could only move along the lines of a grid. Students were encouraged to generate and pursue their own questions about the grid-world, such as questions about the shortest distance between two points in this world. In subsequent sessions, teams were given feedback on their prior work and the work of other teams and were encouraged to continue their work.

Our qualitative analysis aimed at understanding how teams of participants in the VMT online community “bridge” the apparent discontinuity of their interactions (e.g. multiple collaborative sessions, teams and tasks) and exploring the role that such bridging activity plays in their knowledge building over time. We employ ethnomethodological approaches (Garfinkel, 1967) to examine sequences of episodes by using recordings and artifacts from the teams sessions. For our current purposes, we examined 18 team sessions, paying special attention to the sequential unfolding of the sets of four problem-solving episodes in which each team participated. Constant comparison through different instances of bridging in the entire dataset led to our refinement of the structural elements that define bridging activity.

Conclusion and discussion We have identified a set of interactional methods through which participants deal with the discontinuities relevant to their joint activity. Our analysis of this “bridging work” shows that it is a highly interactive achievement of groups and, possibly, very consequential undertaking for their knowledge building over time. In particular, we show how positioning plays a special role in the dynamic way in which teams construct and maintain a joint problem space over time, manage their ongoing participation, and constitute their collective history. Our analysis illustrates the dynamics of positioning in the collective engagement with past and projected work. Furthermore, our analysis suggests that these attempts to establish continuity in collaborative problem solving involve the collective recognition and use of discontinuities or boundaries as resources for interaction (e.g. temporal or episodic discontinuity), changes in the participants’ relative positioning toward each other as a collectivity (e.g. co-narrators and interactive audience), and also the use of particular orientations towards knowledge resources and perspectives (e.g., the problem statement, prior findings, what someone professes to know or remember, etc.). In addition, we suggest that the lens of positioning affords us a more interactive perspective on participation and engagement than the traditional concept of role, at least, within the analysis of long-term, multi-team learning interactions.

References Davies, B., & Harre, R. (1990). Positioning: The discursive production of selves. Journal for the Theory of Social Behaviour, 20, 43-63. Garfinkel, H. (1967). Studies in ethnomethodology. Englewood Cliffs, NJ: Prentice-Hall. Hare, P. (2003). Roles, relationships, and groups in organizations: Some conclusions and recommendations. Small Group Research, 34, 123-154. Renninger, K. A., & Shumar, W. (2002). Community building with and for teachers at the Math Forum. In K. A. Renninger & W. Shumar (Eds.), Building virtual communities (pp. 60-95). Cambridge, UK: Cambridge University Press. Stahl, G. (2005). Group cognition: The collaborative locus of agency in CSCL. In T. Koschmann, D. Suthers & T. W. Chan (Eds.), Computer supported collaborative learning 2005: The next 10 years! (pp. 632-640). Mahwah, NJ: Lawrence Erlbaum Associates. Stahl, G. (2006). Group cognition: Computer support for building collaborative knowledge. Cambridge, MA: MIT Press. Strijbos, J. W., Martens, R. L., Jochems, W. M. G., & Broers, N. J. (2007). The effect of functional roles on perceived group efficiency during computer-supported collaborative learning: A matter of triangulation. Computers in Human Behavior, 23, 353-380.

The dynamics of social roles within a knowledge management community Isa Jahnke University of Bochum Germany isa.jahnke@rub.de

Abstract This contribution presents an empirical investigation into web based communities as an “extension” of a university organization and more specifically into the web based knowledge management processes in the community “Informatics Portal University of Dortmund” (InPUD). InPUD was initiated in order to solve information deficiencies in the field of study management – study planning and practice from the students’ viewpoint – by supporting knowledge sharing between novice and expert students, as well as integrating other formal university roles in the context of study administration. InPUD has an increased number of participants. Since launching on September 2002, more than 1,330 registered participants have written more than 24,100 contributions. InPUD catches today more than 60 percent of computer science students. At the University of Dortmund in the Department of Computer Science there were about 2,000 students enrolled. Registration and login is only necessary when contributing. Reading of contributions is possible without registration and without login; each user can read everything.

The empirical investigation was an action research approach and data was collect with qualitative research methods, which included interviews. The contribution describes the InPUD’s social changes over time and its impact on the organization between 2002 and 2006. In particular, the case study points out the relevance of the roles structure of the dynamics with regard to the successful cultivation of such online communities. It also revealed several important requirements for organizational as well as technical support that should be promoted. Our empirical results will be helpful for web based communities when introduced in large organizations in order to change the social practice in the whole organization. We will explain and discuss these results in more detail.

Extended summary Knowledge sharing processes are activities of individuals who take and play several roles. The notion of the own role, as well as the images of the roles of the others, provides an orientation for role playing which is negotiated formally and/or informally in social interactions. Conflicts in knowledge sharing emerge when role perceptions are incompatible. For example, role owners at universities (students, advisors for students, examination office and registry) have specific expectations about each other role owner, and these expectations may be in conflict. For example in the case of information sharing: when is the right time to convey information, which person needs which information and when should it be available online.

The “Organizational Development of the Computer Science Study” project revealed many information deficiencies at the German University of Dortmund with respect to students’ advisory service, for example the practical realization of (new) examination regulations. The deficiencies were based on different perceptions about varying roles of persons’ own roles and roles of other persons, and led to differences in role making and role playing (Herrmann, Jahnke, & Loser 2004). The mentioned investigation concluded that German students, who decide themselves when, which and how to attend lectures or seminars, lack of information about how to choose their lectures, in which semester and how to plan and manage their studies. The solution included not the reorganization of the role “advisors for students” only, because this role was played by just seven individuals who have responsibilities for more than 2,000 students at the Department of Computer Science at the University of Dortmund.

In 2001, we decided to initiate a web based knowledge management community – “Informatics Portal University of Dortmund” (InPUD) – in order to solve these information deficiencies by supporting knowledge sharing between novice and expert students as well as integrating other formal university roles in the context of study administration. The web based community InPUD combines the knowledge flow among formal roles and their work processes at the mentioned university with other roles (e.g. students/customer). The knowledge sharing in InPUD was voluntary. From 2002 until 2006, we observed the participation as well as social and technical changes of InPUD.

In our presentation we will explain the changes invoked by InPUD’s from the dynamic structures of roles. We will also provide insights into our definition of web based communities (Wenger, McDermott, & Snyder, 2002) and formal as well as informal roles. We will present a detailed description of the “InPUDCommunity” and our research methods and report our empirical results and experiences. Finally, we show what the relevant success factors were for the InPUD-Community from the role perspective.

Method Qualitative study and action research Cultivation as well as evaluation of the web based community InPUD were based on an empirical exploratory study including action research processes (cf. Avison, Lau, Nielsen, & Meyers, 1991), as well as interviews and questionnaires. Our aim was to investigate the reasons for the success of InPUD. We identified empirical based theses, which we term ‘success factors’.

Results Web based communities from the viewpoint of the dynamic of roles.

Success factor 1: Integrating formal and informal roles, as well as primary roles and secondary roles. InPUD is successful because of the integration of both official formal but also informal roles, primary and secondary roles (e.g., students, student’s representatives, student’s advisors and counseling service). Those roles are relevant for a knowledge owner who needs new knowledge and contributes his/her own knowledge for common knowledge sharing, bring formal and informal processes together (Thesis 1).

Success factor 2: Because of the technical system’s ability to bridge spatial and temporal distance, InPUD is successful. The technical system gives the students the chance to participate, to build roles and social networks online. In this manner the technical systems helps to develop social proximity and personal relationships among students. It enables the cultivation of new social relationships and sufficient knowledge sharing for the participants (Thesis 2).

Success factor 3: The cultivation of InPUD is successful due to the fact that participants see the value of usage and benefit of social capital and immediate help (own problems can be solved immediately). Their participation leads to an increase of more social capital. They play a “role” in the development of social relationships and become a part of the community, which can be activated when needed. They are no longer anonymous (Thesis 3).

Success factor 4: Online role presence. InPUD is successful because the formal roles are visible when people communicate online. The online role presence has a positive effect on the development of trust and social proximity. The visible presence of role names improves the assessment of the quality of information (Thesis 4).

The investigation also shows a “moderate facilitation by the formal roles” (delete off-topic contributions, comment wrong answers, clarify which content may discuss, which off-topic is and make it visible). The facilitators had a very moderate position in the community. They did not give questions, but they only answered when other students had no idea or were unable to answer.

Discussion Our empirical results will be helpful for web based communities when introduced in large organizations in order to change the social practice in the whole organization. We will explain and discuss these results in more detail at the conference meeting.

References Avison, D., Lau, F., Neilsen, P. A,. & Myers, M. (1991). Action research. Communications of ACM, 42, 94- 97. Herrmann, Th., Jahnke, I., & Loser, K. U. (2004). The role concept as a basis for designing community systems. In F. Darses, R. Dieng, C.Simone & M. Zackland (Eds.), Cooperative systems design (pp. 163- 178). Amsterdam: IOS Press. Wenger, E., McDermott, R., & Snyder, W. M. (2002). Cultivating communities of practice: A guide to managing knowledge. Boston, MA: Harvard Business School Press.