The transition from surface to meaningful collaboration in online case-based learning: an activity-theory perspective

The use of cases for teaching and learning, while increasingly popular, is a challenging endeavour for both students and instructors. This paper introduces a model which encapsulates the issues involved in online case based learning. The authors argue that, unlike traditional case discussions, students in an online case discussion environment must not only be acquainted with the case method itself, but must also develop productive team relationships. Since this is a complex task even for students in a conventional face-to-face environment, the requirements for meaningful collaboration within a computer mediated communication, where students have little or no experience with the technology and must develop the necessary method and media skills in parallel, can be a time intensive and emotionally draining experience. Our model uses an Activity Theory (AT) perspective to analyse the mediators required during electronic Case-Based Learning (eCBL) and identifies the sub-activities instructors can implement to facilitate meaningful online case discussions. Case-based learning and meaningful participation Case-based learning, first introduced around 1870 at the Harvard Law School, was originally designed to help students build a broader understanding by using real court cases to illustrate legal principles (Merseth, 1991). The concept was then successfully transferred to Harvard’s Graduate School of Business Administration, where cases led to a problem-centred teaching approach in which questions were derived from real life (ibid.). Information Systems (IS) education has rather different requirements, with a major challenge being the integration of technology theory with business analysis – it has not always been possible to see how cases can be used most effectively to supply this need. Gendron and Jarmoszko (2003) have, however, developed a framework for an undergraduate MIS course which shows that the case method is an excellent way to balance hands-on topics with technology theory. In general terms, case-based learning (CBL) claims to develop high-order reasoning skills, to bring real world examples into the classroom, to allow students to learn by doing, to foster constructive and independent thinking; and to bring organisational impacts, social values and ethical issues to the forefront of IS education (Gragg, 1940; Hackney et al., 2003). Reviewing the case method, Dooley and Skinner (1977) made the point that much of the discussion concerning case based teaching and learning was flawed because there is no single, exclusive definition of how to use cases, so that instructors are referring to a multiplicity of pedagogical techniques when they say they use cases. How the method is implemented is strongly related to the instructors’ own pedagogic philosophies, their views on how students learn; and their perception of their responsibility for student activities (Dooley & Skinner, 1977). A second strong impact factor is the case material selected, the case related tasks and the forms of assessment (Reynolds, 1978). If case assessment clearly favours a predefined set of answers, other issues become less relevant (Argyris, 1980). In his influential article ‘Because wisdom can’t be told’ Gragg (1940) compared lectures with case-based teaching, finding that the greatest benefit of the case method is stimulating students’ thinking and collaboration in order to analyse a number of interpretations of a single problem. The focus on collaboratively constructing insight is part of a fundamental distinction between two assumptions about learning: the acquisition metaphor, which views knowledge as a set of concepts readily available for transfer; and the participation metaphor, in which knowledge consists of shared activities, being an object of multi-perspective evaluations (Sfard, 1998). ‘Meaningful participation’ is therefore the true collaborative effort to understand and critically reflect on varying interpretations of a problem, rather than a search for a single, approved solution. The acquisitionist view is deeply rooted in the western understanding and implementation of the educational system and has shaped today’s students’ expectations of what course work should comprise (Sfard, 1998). For us to be able to discuss meaningful participation effectively, we found a major issue to be addressed was the student’s understanding of learning activities which, in addition to quantitative performance measurements (such as number of words, references, or postings), requires the adoption of qualitative measures such as critical evaluation of contributions, dealing with uncertainty and making decisions based on incomplete information. Attempts to analyse or model case based learning (CBL) have thus far been somewhat limited, concentrating on single phenomena such as the design of case scenarios (Ward, 1998), collaborative decision taking (Parent et al., 2002), selection of case material (Somervell et al., 2004), or the structuring of case material (Dabbagh & Denisar, 2005). A more comprehensive approach including the analysis of cognitive, social, teaching and discourse processes was undertaken by Heckman & Annabi (2005) – though these authors have limited their model to asynchronous, text-based learning interactions. Jonassen & Hernandez-Serrano (2002) discuss the creation of content libraries in case-based reasoning; and Jarz et al. (1997) describe a virtual environment where learners navigate through multimedia enhanced cases. Both these models, however, focus on the presentation of the case material. Our model is concerned with collaborative electronic case-based learning (e-CBL), which requires instructional designers to: create opportunities for online communication; formulate tasks which lend themselves to discussion; and support learners in their media choice and group management tasks. Focusing on collaborative and mediating effects, the model provides a different lens for analysing online CBL by defining learning activities as central design criteria. The theoretical perspective we have chosen is based on social constructivism and stresses the importance of shared meaning construction, multiple perspectives on information, and learners’ engagement in guiding their own learning (Jonassen, 1999). In this paper we initially present a perspective of case-based learning based upon Activity Theory, adapting Engeström’s classic diagram to the instructor’s and learner’s perspective and showing how this enriches our understanding of the social and contextual aspects of CBL. We then introduce our own theoretical model of the online CBL environment, discussing the role of artefact mediation in the learning process and proposing a view of how this might be used to develop a model of online CBL based on Engeström’s activity systems. An Activity Theory view of case discussions To enable the development of a broader model, capable of integrating the various findings from CBLrelated research, we took an activity-theoretical approach to the analysis and design of case-based learning. Activity Theory (AT) states that acting is a precursor to learning, thus supporting constructivist and social views of learning, as opposed to the cognitive approach which emphasises the development of the individual through instruction (Barab et al., 2003; Jonassen & Rohrer-Murphy, 1999). Our use of a social perspective for our model development does not devalue the cognitive research approach (Anderson et al., 2000), but is simply the consequence of choosing a broader analytical context as a result of observing learners in situ (Nardi, 1996). Central elements of Activity Theory Based on an analysis of the literature Lazarev (2004) identifies the following main characteristics of activities which are the primary unit of analysis in AT: (1) activities define humans’ relationship with their environment; (2) activity implies free design and choice of goals; (3) activities are shaped by socio-cultural programs which are themselves the product of activities; and (4) activities are open to development and are primarily accomplished in conjunction with others. Applying the concept of activities to learning brings the agency of learners (i.e. their responsibility for their own learning) to the forefront of the discussion about what can and can not be achieved by instructional design. Repkin (2003) questions the effectiveness of today’s educational system where more and more courses are developed in response to an ever-growing demand for highly skilled employees. Interpreting learning as a goaloriented, contextualised activity provides a new perspective on education. Applied to CBL, goal-orientation substitutes for the memorising of content pre-specified by the instructor, answers to collaboratively identified issues in a particular case where the application of knowledge is simultaneously put into context and, more importantly, generated from analysing that context. Here learners ultimately determine their own learning needs to understand the situation presented in the case. But, as Repkin (2003) clearly states, not every activity is learning, only those activities which provoke a change in the agent. Although it is not always easily discernible, activities which merely strive to reproduce quantitative conditions for a good mark (e.g. meeting the required number of pages) would appear to imply less internal change than activities attempting to make sense out of the issues identified. The contradiction of activity goals is described by Engeström (1987) as the primary contradiction of commoditization: the conflict between ‘use value’ and ‘exchange value’. Engeström vividly evokes Leontiev’s village physician who, in healing people, not only decreases his source of income but may also be asked to upgrade his practice to meet his patients’ expectations of more up-to-date treatment! There are clear parallels with the situation facing students in a case-based learning environment, who can influence both their learning content as well as its amount. Since time is a limited resource, these students must often choose between getting the case done and following up an unresolved case issue. In terms of use and exchange value this means they have to balance their time between the minimal effort required to pass the course and get a grade in exchange; and the investment of time required to really learn from the course, which would provoke a change of understanding. Case-based learning as an activity system Engeström’s structure of two superimposed triangles, illustrated in Figure 1, is a widely-used visualisation of an activity system. The inner triangle, pointing downwards, shows how the individual interacts with his/her environment through community rules defining the ‘being together’ and the ‘doing together’ (Engeström, 1987). The outer triangle, pointing upwards, introduces the goal-directed usage of tools and regulations specifying human activity. ‘Being and doing together’ develops into explicit traditions and regulations about social interaction (any form of communication) and the distribution of tasks (including potential rewards). The final addition to the structure is intentional usage of tools, which Engeström derives from Vygotsky’s idea of mediated memories and mediated activities. Figure 1: The human activity system (Engeström, 1987) A simple memory aid, such as a notched stick for a signpost, can be seen as an externalised part of the human natural memory. This allows experiences to be coded as signs, which is a first step toward the shared construction of meaning . The product of mediated memories is self-generated, artificial stimuli which influence human behaviour (Vygotsky, 1978). This enables activities to be mediated, not only through physical tools, but also through cognitive tools (signs, programs). Figure 2. Case-based learning activity systems: design and learning perspective, based on Engeström (1987) Figure 2 shows the application of Engeström’s activity systems to e-CBL activities. In understanding these two diagrams, it is important to realise that instructor and student views are both part of a third activity system(not displayed), that of case discussions in online learning communities. For the purpose of designing and managing the learning environment, however, it is helpful to separate the learner and designer perspectives (Barab et al., 2004). The left-hand part of figure 1 shows an instructor who is creating a learning environment to facilitate CBL for external students. As a member of a school and a university the agent follows explicit rules as well as unspoken norms. This is further explained by Giddens’ characterisation of interdependent relationships between individuals and groups as recurrent social practices, regulated by rules and resources belonging to the group (Giddens, 1979). These rules may concern the way examinations are taken or selection of the primary medium for exchanging messages. Not all activities are the result of thoughtful planning – some activities happen simply because this is the way things are usually done within that particular School. At this level not all rules have direct regulatory consequences and not all rules need explicit formulation: “ ‘To know how to go on’ is not necessarily, or normally, to be able to formulate clearly what the rules are.” (Giddens, 1979, p.67). This is particularly true if the goal of the activity system is analysed in terms of use value (student learning) and exchange value (student feedback given to the instructor). The way the instructor steps up the career ladder can be directly linked to student feedback about his or her teaching. CBL is an innovative approach to learning which confronts students with the uncomfortable notion of uncertainty, because no single pathway to the ‘right’ solution is indicated. Exchange and use values need to be brought into alignment, for example by stressing the fact that companies ask for more than excellent grades, and thus students who have learned to solve problems collaboratively are better prepared to cope with the dynamic nature of a future workplace (Pillay, 2002). Engeström compares the outcome of “school-going” (reproducing solutions for well-structured problems) with what pupils are expected to do in an ever more complex work activity system (Engeström, 1987). To differentiate mediation through community membership (the specific case for e-CBL) from mediation through ‘division of labour’ (the general case in activity theory), it is important to understand the origin of both types of mediators. Community rules describe the modus vivendi of a group and exist to ensure the continuity of the group, whereas task distribution rules exist to ensure that the group’s common objective is achieved ((Engeström, 1987). Applied to the e-CBL scenario, division of labour would include tasks such as organising group meetings, searching out specific information for the case, or determining whose turn it is to write up and submit the case. Although not always clearly distinguishable, community rules would describe when and how free riders are reported. It is beyond the scope of this paper to describe in detail all possible relationships for both activity systems presented. Examples, although not related to CBL, for the instructor’s view can be found in Jonassen & RohrerMurphy (1999) and the learners’ view is analysed in Collis & Margaryan (2004). The third mediator in Engeström’s analysis of activities is technological and cognitive artefacts (tools). In the literature of educational technology the impact of “new” media on distance education already has a substantial history. Every new medium (film, television, radio) has claimed to improve learning, but providing solid evidence of educational success over a number of studies has been problematic (Clark, 2001). Comparative studies concerning the influence of media on learning have not been able to show unequivocal results (IDECC, 2004a, 2004b) until the question was reframed from “Do media influence learning?” (Clark, 1983) to “Will media influence learning?” (Kozma, 1994). This suggests that educational technology is a design science (Simon, 1996) rather than a natural science – that is, if there is no apparent relationship between media and learning efficiency, this is only because the link has not yet been made (Kozma, 1994). Reading the Kozma vs. Clark debate from a activitytheoretical perspective, we take the view that media are not mere instruments of delivery, but rather mediators of instruction which themselves favour or impede learning when placed in a specific context with concrete learning objectives and learning conditions (community and distribution rules). The following section elaborates the role of artefact mediation during the learning process and explains how it can be used to develop a model of online CBL based on Engeström’s activity systems. Cognitive and technological mediation of case discussions In this section we introduce our theoretical design framework for online CBL while transforming the activity-system ‘online case discussion’ into a process-model allowing for the adoption of design recommendations from cognate research areas such as critical thinking (Astleitner, 2002), group theory (McGrath, 1991), media synchronicity theory (Dennis & Valacich, 1999), channel expansion theory (Carlson & Zmud, 1999) and computer supported social networks (Garton et al., 1997). As outlined in the previous section, the concept of activity systems provides a lens for analysing specific activities within the context in which they occur. To obtain a process model we take the three mediation categories: tools, community and division of labour (Engeström, 1987), identify potential problems originating from mediation and refer to other research areas where similar problems have been investigated. This process consists of various support activities initiated by the instructor or the system in order to facilitate mastery of the very skills which are pursued by online CBL. In order to produce any kind of development, activity-systems need to overcome contradictions, which can exist either between the required and existing skills needed to solve a problem, or between the required and existing tools needed to exchange information effectively. Engeström(1987) suggests a hierarchical structure of contradictions, where higher level contradictions occur within a mediator (students question the usefulness of collaborative learning) or between mediators (existing group norms oppose collaborative learning). At a lower, operational level these contradictions are embedded in a net of parallel activities – for example, over time the group develops rules governing how to prepare effectively for a group meeting, or how to make efficient use of a communication medium (in Engeström’s nomenclature those would be rule-producing and subject-producing activities). However, while parallel activities can remedy a contradiction they can also exacerbate a problem at a higher level, e.g. the development of online communication skills may shorten the time needed to produce a good case report and thus lower the exchange value of this learning activity. The process model is designed to encapsulate recurrent problems in online case-based learning and to formulate instructional interventions (Kuuti, 1991) which enable learners to overcome barriers during the execution of their learning activities. The different stages of development are derived from an extensive literature review, focusing particularly on problem-based and collaborative online learning, as well as upon ongoing observations of teaching practice (Voigt & Swatman, 2005). Figure 3 : Reaching insightful participation The ultimate goal of the process model shown in figure 3 is a meaningful case discussion, the importance of which was discussed in the first section of this article. Starting with the individual learner the instructional interventions develop along two paths: quadrants I and II address the learner’s own methodological capabilities, while quadrants III and IV show the facilitation of necessary media competencies. Those two lines correspond to Engeström’s activity mediation through cognitive, social and artefactual tools (Engeström, 1987) A second goal of the model is the learner’s development as an individual as well as a team member, which again takes up the idea of activities being shaped through task-distribution and community rules. This idea goes back to Vygotsky’s concept of a Zone of proximal development (ZPD) and the importance of shared experiences for learning. ZPD describes the ‘the distance between the actual developmental level as determined by independent problem solving and the level of potential development as determined through problem solving under adult guidance or in collaboration with more capable peers’ (Vygotsky, 1978, p.117). The ‘embeddedness’ of individuals in their group context is depicted through bidirectional arrows between quadrants I and II and quadrants III and IV. It is important to note that although the process model builds upon the learner’s existing knowledge, the separation between the individual and the group is dialectical rather than chronological. There is no assumption of stand-alone learners accumulating knowledge before they enter the group work arena. We agree with Sfard (1998), however, that there is value in integrating parts of the acquisitionist view into the ‘learning-as-participation’ metaphor. Learners go through the process shown in figure 3 several times during each case they tackle. There are 5-6 cases in the course which formed the development environment for the model, where students participate in discussion processes, actively applying and testing their knowledge (participation metaphor). While there are similarities in terms of the learning cycle, however, each case is designed to provoke changes within the learner which, independently of the situation, enables improved participation in the next case (acquisition metaphor). Cognitive mediation: critical thinking and the construction of shared meaning (Q I and II) Quadrants I and II describe the development of case discussions as cognitively mediated activities, first discussing the issues and then suggesting instructional intervention possibilities to resolve some of these issues. Issues (Q I): The instructor who releases the case (plus some additional readings) may do so by saying: ‘Please identify the salient issues in the case, discuss them in your group and don’t forget to give justifications for the solutions you come up with.’ At this moment the instructor is mediating student activities, based on what s/he assumes to be a fruitful approach to the case-based learning. Even if students have undergone a theoretical introduction to what comprises the case-method, however, practice will still give rise to new questions which are not covered by the available guidelines. To avoid simplistic and superficial interpretations of cases, critical thinking capabilities must be facilitated to shift students’ attitude from passive consumers to active problem solvers (Tan, 2004). A tentative definition by Paul (1993, p.7) outlines two characteristics to differentiate critical from basic thinking: (a) critical thinking entails self-improvement and (b) self-improvement is caused by evaluating thinking against intellectual standards relating to relevance, accuracy, fairness and completeness (among others). Applying those standards can be a challenging task for first year, usually younger students (Tolhurst & Baker, 2003). There are numerous studies which show that students’ epistemological beliefs influence their active involvement, comprehension and persistence in learning (Schommer, 1994) and these beliefs are also relevant to students’ attitude towards their peers’ opinions and, consequently, to their appreciation of group discussions. Cases allow for multiple interpretations of facts yet, alongside intellectual standards, students also need to develop intellectual virtues – to become fair-minded thinkers, able to recognise their own biases and limits (Paul, 1993: p.258). Instructional intervention: Research suggests that providing models about critical thinking alone does not yield the desired effect and concrete monitoring and scaffolding activities are needed to improve students’ critical thinking skills (Astleitner, 2002). Scaffolding is closely related to Vygotsky’s ZPD and aims (finally) at independent task mastery by the student (Cheyne & Tarulli, 1999). Regarding a definition of the nature of scaffolds, Pea (2004) pointed out that the term has become overly inclusive and, returning to the original definition from Wood (1976), he suggested that scaffolds are those properties of a learning environment which channel, focus or model learning activities and, most importantly, can be calibrated according to the learners’ actual ZPD. The point with technological scaffolds, however, is that their automated application needs still further elaboration in order to react to continuously changing learner needs (Pea, 2004). Although the concept of scaffolds is not exclusive to quadrant I, we see their most effective application at the beginning of the case discussion where they can enhance students’ qualitative understanding of the task and thus create a need for further collaboration (Astleitner, 2002). Issues (Q II): The requirement to discuss something meaningfully demands that the group develop a shared understanding of what is meant by ‘meaningful discussion’. The problem of misunderstanding in communication is also deeply related to the problem of limited mutual knowledge (Cramton, 2001), so that neither the instructor nor the students have much understanding of their interaction partner’s broader or immediate context, e.g. whether someone has any previous experience with the case method, or whether technical facilities at the other end of the communication channel are present (Internet connection, available software packages, etc.). Pearce and Cronen (1980) suggest that the construction of shared meanings happens according to a hierarchical frame of different context layers including the actors’ cultural background, their self-perception, and the relationship between the actors and their specific communication goal (speech acts). It is beyond the scope of this paper to discuss these authors’ theory of coordinated management of meaning (CMM) in detail, but the notion of hierarchically organised meaning construction suggests that issues need to be addressed at different levels of communication. A learning group comprising students of differing nationality may need scaffolds to foster the inclusion of all team members, or instructor intervention can help to create awareness of different communication habits (Shih & Cifuentes, 2003). Similarly, contradictions between group members may arise due to different propensities to define a relationships: what is included in the ‘contract’ and what are the criteria for success and ending of the contract (Pearce & Cronen, 1980). A huge part of group research is dedicated to the factors influencing successful relationships within groups and refers to phenomena like ‘free riders’, ‘social loafers’ or ‘suckers’ as social dilemmas, where “the rational pursuit of self-interest can result in collective disaster” (Kerr, 1983). Research on how to avoid those effects in unstructured tasks (Salomon & Globerson, 1989) such as case discussion, however, is still limited. Instructional intervention: According to Smith (1998) many problems of uneven workload distribution can be eradicated through careful group formation, task structuring and formative assessment. One way to structure free collaboration in computer supported collaborative learning (CSCL) is scripts, which describe explicitly the details of a contract between instructor and students about how they should collaborate and solve emerging problems (Dillenbourg, 2002). The impact of scripts on higher-order learning is yet to be shown, however – it seems as though there is a trade-off between using a script to engage all team members equally (Miao et al., 2000) and using them less to avoid the impression of excessive coercion, which might decrease student motivation (Dillenbourg, 2002). Over-scripting of collaboration in e-CBL can jeopardise a core requirement of the case method – the critical evaluation of ‘quick’ answers which leap first to mind. Scripts can distract students from the real effort, which is to review a complex situation from a number of different angles, although scripts can help to reduce uncertainty about what to do next in the sometimes daunting complexity emerging out of the case itself and the communication technology used (Tan, 2004). Scripts for e-CBL can be found in any form of work-through guideline for how to approach a case, such as the Harvard Business School method (Christensen & Hansen, 1987) or the more recent MICA method (Siciliano & McAleer, 1997). Technological mediation: affordance and collective choice of communication tools (Q III and IV) According to Jonassen (1999) there are three main components defining a constructivist learning environment: rich context description, authentic problem representation, and an engaging problem manipulation space. Jonassen later differentiates between representation/manipulation tools which mediate activities between the learner and the technology problem; and conversation/collaboration tools which mediate interactions in communities of learners (Jonassen, 1999). Quadrants III and IV represent largely instructional interventions which improve learners’ ability to present their ideas (presentation space) and to negotiate a common understanding of one another’s contributions (manipulation space). The importance of viewing the lower two quadrants in conjunction with the upper two quadrants cannot be over-stressed – many issues identified in Q I and Q II can and should be addressed through cognitive and technological mediators simultaneously. Issues (Q III): Expecting students to communicate online the same way they would in a ‘real’ encounter can lead to little or no interaction if the learning environment does not match the specific human dialogue conditions in a virtual context (Sorensen, 2003). Wellman et al. (1996) note that: “Computer supported social networks are developing norms and structures of their own. They are not just pale imitations of ‘real life’. The Net is the Net.”. This paragraph analyses the meta-cognitive dimension of the online collaboration, addressing the problem of the more limited experiences available in electronic communication media fairly briefly. For those seeking further clarification, we refer to earlier work (Voigt & Swatman, 2005) where we used the example of synchronous text communication (chat), finding that in an online course where use of chat was voluntary , only 2 out of 16 teams engaged in substantial regular online meetings. Reasons why students neglected chat communication ranged from technical difficulties in starting the chat application to a lack of commitment and preparations for the planned chat sessions. Despite the great potential available from synchronous meetings, successful experiences were an exception and most teams abandoned chat in preference to forums which appeared to offer comparable benefit with less effort. The choice of a communication tool is related to our perception of tools in terms of what they have to offer or, as Gibson’s phrases it, people relate possible actions to the objects they see and this is what characterises the affordance of an object (Gibson, 1979). Affordances of artefacts are subject to learning: whether an object suggests any activity or not depends as much on how it is perceived as it does on the experience and interpretation applied to the perception (Bærentsen & Trettvik, 2002). Instructional interventions: The importance of promoting in-depth knowledge of communication tools at an early stage is supported by McGrath’s (1991) theory of groups. McGrath differentiates 4 different group activity modes: (a) inception mode (selecting a work strategy, choosing interaction and inclusion opportunities), (b) technical problem solving mode (selecting working procedures, defining roles and policies for task and reward distribution), (c) conflict resolution mode (resolving differing values and preferences, managing power and payoff allocation as well as the re-negotiation of those allocations) and (d) execution mode (carrying out the concrete tasks). Tip (????) suggests that only the inception and the execution mode are a necessary part of any group project and that groups tend to avoid technical and political problem solving activities, if possible, since group members’ resources are limited and therefore members try to minimise their efforts. Quadrant III therefore suggests that instructors introduce early experiences with the various communication technologies outside the pressure of a fullblown case discussion. To provide positive experiences we recommend mini-encounters, where small tasks are approached using a number of different media. In accordance with McGrath’s group theory, those tasks could includeL simple meeting coordination, case discussion or an informal getting to know each other. To ensure these media training sessions do provide the intended experience, existing research matching tasks and media attributes is taken into account. It is not possible to analyse all task-media combinations within our model, but the advantages of synchronously shared learning spaces (such as chat or a group editor) can be seen from the shared understanding obtained by capturing group cognition in the form of a list of issues or a mind-map displaying the case relationships (Stoyanova & Kommers, 2002). Another advantage of synchronous communication is its ability to deliver immediate feedback, which improves information processing for more equivocal tasks (Dennis & Valacich, 1999). Equivocality refers to multiple interpretations of ambiguous information as they are usually given in case studies. And borrowing from conversational analysis research: peers construct shared meanings incrementally “through cycles of displaying, confirming, and repairing shared meanings” (Roschelle, 1992). An efficient way to repair divergences are iterative turn-taking structures such as chats, where students can build on each others’ utterances to converge their understanding. Grasping how media influence communication in specific task-media combinations has proven crucial for students’ ability to enter and participate in online discussions (Bregman & Haythornthwaite, 2003). Issues (Q IV): Quadrant IV addresses the impact of aggregated media usage patterns within a group on overall group efficiency and the individual’s appropriation of media. Both issues can be explained by the research area of computer supported social networks (CSSN) (Wellman, 1996). In contrast to the focus on the individual in quadrant III, in quadrant IV we adopt a network perspective, highlighting the importance of transactions between individuals In this context Haythornthwaite (2001) explains that communication media fulfil more than a purely work related exchange of information and suggests that parallel advice, socialising and emotional networks also need to be supported. To measure an individual’s integration within a group, CSSN analysis offers the concept of local and global centrality (Scott, 1991). Locally central members would be highly connected students within their learning group. Global centrality would analyse a member’s connectedness at a class level where students may maintain relationships with members of other groups. Despite being part of a relatively inactive group, individuals could still enjoy centrality in a broader forum, where non-group tasks are discussed. Haythornthwaite’s (2001) study indicates that, over time, activities move from class-wide to group-based interactions, increasing the isolation of learners with higher global and lower local centrality within their emotional support network. Low logistical and social costs of participating in CSSN and the egalitarian nature of net communication encourage reciprocal supportiveness, although the ease of initiating online relationships is opposed by individuals’ flexibility in switching between groups of ties, which allows them to drop any engagement without providing reasons to the “abandoned group” (Garton et al., 1997). Maintaining community ties involves time and “emotional resources”. Both are limited resources for learners and ties from online nets compete for these resources with demands from the off-line world (Kazmer & Haythornthwaite, 2001). A second issue addresses media appropriation within groups as it is related to group effectiveness. Markus (1987) sees the introduction of interactive media in groups facing the same problems innovations usually have to overcome during their diffusion phase: gaining ‘universal access’ and simultaneously retaining early adopters where reciprocity from communication partners is lacking. Universal access within a group means that each member has access to and can be reach through the introduced medium (Rogers, 1995). Reciprocity refers to the dependency of early adopters on the adoption rate within their group (Markus, 1987). Although an individual may be convinced on logical grounds (benefits) to discuss a case in a chat session rather than in a forum, the medium is of little help if no other group member joins in. Over time groups develop communication routines and changes at a later stage are difficult to implement (Huysman et al., 2003; Tyre & Orlikowski, 1994), even if an increasing degree of case complexity would suggest it was logical to do so. According to Markus (1987), a group is unlikely to maintain multiple communication modes for a single purpose. In order to include all members the lowest common denominator will be chosen in the long run to minimise information distribution costs, even where the chosen medium does not adequately support all the necessary group functions suggested by McGrath’s (1991) theory of groups. Instructional interventions: If interventions in quadrant III were primarily aimed at providing learners with media experience, those of quadrant IV are designed to provide a positive, mediated group experience. To avoid ‘media-frustrated’ groups, instructors need to address the issues of media-based member integration and technology diffusion. Connectedness is an important precondition for exchanging resources, yet the shrinking of connections outside the learning group is partially caused by task design which requires no out-bound communication (Haythornthwaite, 2001). We therefore suggest that supportive discussions be held in a class-wide forum, so that connections can be maintained outside groups without starting an additional stream of tasks. Problems which originate in a particular group-forum could be formulated by the instructor as a general problem statement to be discussed in the class-forum once all teams have submitted their reports. Discussing their problems in a public space requires students to understand that there is no competitive condition between the individual teams and that their group’s exposure serves the objective of preventing mistakes already made by other groups. A study conducted by Constant et al. (1994) showed that people are more inclined to share advice than tangible information, because advice relates to people’s need for self-expression, while their understanding of whether the information belongs primarily to their group or the class affects willingness to share. Attitudes towards the sharing of information are closely linked with the norms learners experience within their context – specifically, whether the class gives visible credit to their contributions or not. In the end, information sharing has not only an informing value for decision making but also a symbolic value for status attainment (Constant et al., 1994). To foster the adoption of communication media, we suggest promotion activities (e.g. instructors can offer synchronous events where students obtain exclusive information) or using the medium’s ‘play value’ (Markus, 1987), where students can engage in informal chats. Although mostly brief and opportunistic informal communication can play an important rule in organising more complex activities, such as if a group member needs an interim input to continue with his/her task, or if an additional online meeting seems necessary (Whittaker et al., 1994). A means to provide technological support for informal communication could be the usage of Instant Messengers (IM), which support a sense of proximity; and the indication of online status (back in a minute, please don’t interrupt, etc) shows other people looking for an “informal exchange” when to launch a question and when not. Managing the obtrusive side of incoming messages is an important feature to keep the tools effective, however (Nardi et al., 2000). Conclusions and outlook This paper has endeavoured to answer to the question: ‘How can instructors facilitate the emergence of meaningful collaboration?’. We began by outlining what makes participation meaningful within the context of online case discussions. We then analysed case discussions from an activity-theoretical perspective, noting that case discussions resemble activity systems as described by Engeström (1987) by having the following characteristics: (a) they comprise goal directed activities, (b) those activities are mediated by cognitive and technological tools as well as by community and distribution rules, (c) most activities are accomplished collaboratively. Activity systems was a helpful lens (a) to analyse learning activities for individuals and in a group context and (b) to identify possible sources of tensions during the learning and discussion process. Based on those two categories of findings the authors developed a process model presenting the trajectory of students during case discussions (figure 3). For analytical purposes student activities are split into learning the case method and mastering their communication tools. A second, horizontal split in the model addresses learner development through transactions with their group environment (Vygotsky, 1978). For each of the four resulting quadrants we have developed a number of issues representing contradiction within the overall activity system. The appearance of an issue in one quadrant does not mean that influencing factors can not be found in other quadrants, reflecting the ‘nested nature’ (Jonassen & Rohrer-Murphy, 1999) of activities within the process model. Based on a broad literature review and personal observations of online CBL, we then suggested a number of instructional interventions to mitigate tensions during the collaborative effort to master the case and the technology simultaneously. Although we believe that these identified issues and suggested interventions are helpful, they are not sufficient for all forms of CBL under all conditions. The process model as presented, however, is an example of how activity theory can be used to inform the design and management of a collaborative learning environment. Quadrants I and II of the model covers the areas of critical thinking (Paul, 1993) and efficient teams (Kerr, 1983), which clearly pre-date web-based education. This reflects our belief that online education is as much about education as it is about new media technologies. In order to get the full benefit of the online environment, much effort is needed to organise case discussions on a procedural level. Just as collaboration does not always work in the off-line world if not structured appropriately (Salomon, 1992; Salomon & Globerson, 1989), simply putting new media at students’ disposal is no guaranty for successful collaboration (Rice, 1999).