Wordt gepubliceerd in: P. A. Kirschner, S. J. Buckingham-Shum & C. S. Carr (Eds.), Visualising argumentation: Software Tools for Collaborative and Educational Sense-making (pp. 51-73). London: Springer. ISBN 1-85233-6641-1
3 esigning Argumentation Tools for Collaborative Learning
Gellof Kanselaar, Gijsbert Erkens, Jerry Andriessen, Maaike Prangsma, Arja Veerman and Jos Jaspers
Department of Educational Sciences, Utrecht University, the Netherlands
TNO Human Factors, the Netherlands
The focus of education has shifted towards working actively, constructively and collaboratively, as this is believed to enhance learning. The studies discussed here deals with the influence of different CMC (Computer Mediated Communication) tools on argumentation processes during collaboration. The purpose of our research is to investigate the effect of computer supported environments and its tools on the final product through differences in the participants’ collaboration processes. In this chapter we will concentrate on students collaboratively taking part in argumentation via CMC systems. Computer environments that support collaborative writing can emphasize both the constructivist and collaborative aspects through its active and interactive nature.
3.2 Argumentation and Collaboration in CMC Systems
One of the main principles of constructivist learning theory is the negotiated construction of knowledge through dialogue. Such learning through negotiation can consist of testing understanding and ideas against each other as a mechanism for enriching, interweaving and expanding understanding of particular phenomena. Active engagement in collaborative argumentation during problem solving fits this principle by giving prominence to conflict and query as mechanisms for enriching, combining and expanding understanding of problems that have to be solved (Savery & Duffy, 1995). After all, as Von Glaserfeld (1989) has noted, other people are the greatest source of alternative views to challenge our current views and hence to serve as the source of cognitive conflict that stimulates learning.
Knowledge is actively constructed, connected to the individual's cognitive repertoire and to a broader, often team-based and interdisciplinary context in which learning activities take place (Salomon, 1997). Constructivism seems to be influenced not only by a Piagetian perspective on individual cognitive development through socio-cognitive conflict, but also by the socio-cultural approach emphasising the process of interactive knowledge construction in which appropriation of meaning through negotiation plays a central role (Greeno, 1997). From a constructivist perspective, collaborative argumentation during problem solving can be regarded as an activity encouraging learning through mechanisms such as externalising knowledge and opinions, self-explanation, reflecting on each other's information and reconstructing knowledge through critical discussion (Kanselaar, de Jong, Andriessen, & Goodyear, 2000; Kanselaar, & Erkens, 1996).
We consider an argument to be a structured connection of claims, evidence and rebuttals. A minimal argument is a claim for which at least doubt or disbelief is expressed (van Eemeren, Grootendorst & Snoeck Henkemans, 1995). Such doubt or disbelief can be expressed by an individual (if working alone) or by a partner in an argumentative dialogue. In response to such doubts a complex structure may be produced potentially including features such as chaining of arguments, qualifications, contraindications, counter-arguments and rebuttals. Hence the argument is the product, the structure linking claims, the evidence or rebuttals. The process by which the argument is produced we refer to as argumentation.
Our interest lies in argumentation structures that are built by groups of students involved in collaborative problem solving and writing. During problem solving we expect students to make various claims about the domain and the potential solutions. It is possible that during the problem solving no doubt is expressed regarding claims and solutions and hence no argument emerges in the dialogue. However, such a situation seems unlikely and we believe would not produce the best solution to the problem. Certainly if the students have not produced reasons to support the claims and solutions during the problem solving process itself then we have no reason to believe that they will be able to produce such reasons at a later date. Therefore we believe that students should be encouraged to use argumentation processes to build argument structures during problem solving.
We will concentrate on students collaborating via computer mediated communication (CMC) systems. Communicative tools give access to collaborating partners through Computer Mediated Communication (CMC) facilities like chat and discussion forums, but also to other resources, such as external experts, or information sources on the Internet. In this respect, the program functions as a communication medium (Henri, 1995). The collaborative aspect is mainly realized by offering computerized tools that can be helpful for collaborating students in solving the task at hand (e.g., the CSILE program of Scardamalia, Bereiter & Lamon, 1994; the Belvédère program of Suthers, Weiner, Connelly & Paolucci, 1995). These tools are generally one of two types: task related or communicative. Task related tools support task performance and the problem solving process (Roschelle & Teasley, 1995; Salomon, 1993; Teasley & Roschelle, 1993). Programs that integrate both tool types are generally known as groupware: they are designed to support collaborative group work by sharing tools and resources between group members, and by offering communication opportunities within the group and with the external world.
This chapter addresses how argumentation processes can be supported in electronic environments. We studied students actively engaged in collaborative argumentation in order to solve open-ended problems such as writing argumentative texts, constructing hypotheses or designing computer-based learning programs. These types of problems are characterised by the existence of justifiable beliefs and multiple acceptable viewpoints, as described by Baker (1992), Andriessen, Baker, & Suthers (in press). In working on problems together, students first have to establish a (partially) shared focus, which can be changed, maintained or refined during the problem solving process (Roschelle, 1992). The focus determines the concentration on thematic parts (sub-problems) of the problem to be solved. Subsequently, information relevant to the sub-problem must be generated and gathered from mental or material resources. The next phase is to critically check its strength (Is the information true?) and relevance (Is the information appropriate?) before integrating it in the problem-solving process (for instance by assimilating new information in a writing assignment). Finally, after discussing alternative solutions the strongest and most relevant one must be chosen (Erkens, 1997).
3.2.1 Interface Design for Argumentation
To provoke and support argumentation in CMC systems, interaction can be structured at the interface. Dependent on task characteristics, students can be provided with dialogue markers, sentence openers and turn taking control (Veerman & Andriessen, 1997; Veerman, 2000; Veerman, Andriessen, & Kanselaar, 2000). These options might improve shared understanding, focus maintenance or critical assessment of new information. Additional options for free text interaction could stimulate elaboration whereas careful use of turn-taking control and dialogue rules could guide the interaction without constraining it. In addition, graphic representation of arguments might support exploration of multiple perspectives and identification of misconceptions and gaps.
Veerman & Treasure-Jones (1999) studied how to provoke and support argumentation in electronic collaborative problem-solving situations, considering the cognitive processes of critical information checking, argument elaboration and the taking of multiple perspectives. In addition, maintenance of focus was discussed as an important factor in effective argumentation and collaborative problem solving. Five studies on different CMC systems were reviewed, which were all designed for educational tasks and in which argumentation was emphasised as a method for collaborative problem solving or an end goal for learning. The selected CMC systems demonstrated a range of approaches to structuring interaction at the user interface in order to support communication, and more specifically, argumentation (e.g. turn-taking control, menu-based dialogue buttons, and graphical argument structures). In discussing the success of the systems at provoking and supporting argumentation, characteristics of the task, instruction and structured interaction were considered. The review revealed that structuring interaction at the interface does not necessarily provoke argumentation. Rather, the initiation of argument seems to be related to task characteristics, such as the use of competitive task design. However, providing a combination of structured and unstructured interaction modes may support argumentative processes. In communication windows (chat boxes), combining free text entry with well designed argument moves or sentence openers can stimulate students to critically check information. In task windows constructing argumentative diagrams can improve the exploration of multiple perspective taking and argument elaboration. However, some task characteristics can also enhance such processes. Therefore, task features and structured interaction at the user-interface must be considered in close relationship to each other in order to support argumentation in CSCL situations. In addition, offering support for focus maintenance was proposed as an important factor.
3.2.2 Argumentation in NetMeeting, Belvédère and Allaire Forums
In line with this research, three experimental studies were subsequently organised that examined student groups’ academic discussions mediated by the synchronous CMC systems NetMeeting and Belvédère, and the asynchronous system Allaire Forums (Veerman, 2000).
Figure 3.1: Screendump of the Belvédère system.
All discussions were analysed and compared on the factors of focusing, argumentation and the production of constructive activities, a measure that was used to define collaborative learning-in-process. In addition, various forms of pedagogical support were considered, provided by humans or the user-interface.
Figure 3.2: Screen dump of a discussion ‘thread’ in Allaire Forums.
The results can be summarised under the following three headings:
The results showed, first of all, that a study of collaborative learning from electronic discussions requires analyses of focus in relationship to argumentation. Constructive discussions were particularly focused on the meaning of concepts, and included focus shifts back and forth between the application of concepts, while information was critically checked.
Second, ‘indirect’ forms of argumentation in particular were shown to be effective (i.e. checks, mainly by verification questions), in contrast to ‘direct’ forms of argumentation (challenges, counter-argumentation). The more information was checked, the more constructive activities were produced. Absent effects of the ‘direct’ forms of argumentation were explained by the use of the following paradox: to engage in critical debate, students should have well-established views on subjects and be able to mutually recognise opposed knowledge and attitudes - (Baker, De Vries & Lund, 1999). However, a characteristic of the knowledge of students engaged in debates for collaborative learning purposes is that these views are not always well elaborated, since they are subject to the learning process.
Third, the discussions mainly contained additions, explanations and evaluations. Summaries or information transformations hardly occurred. This was due to the cognitive effort required, but also to an incomplete, intuitive and personalised understanding of information under discussion (Kuhn, 1991). To transform information, there must be a certain level of (shared) understanding. In the studies considering task characteristics, students’ preparation activities, prior knowledge and time available for discussion, obtaining (deeper) understanding may have been the highest goal achievable. Reaching new insights may have been just the next step, for instance, when students were sufficiently prepared and had established a mutual framework for interpreting each other’s information in order to engage in critical, hefty discussions (Coirier, Andriessen & Chanquoy, 1999).
Discussions mediated by the synchronous CMC systems NetMeeting and Belvédère and the asynchronous CMC system Allaire Forums, appeared to have different characteristics concerning focusing, argumentation and the production of constructive activities. Relatively speaking, the synchronous discussions in NetMeeting and Belvédère included more ‘direct’ forms of argumentation (challenges, counter-argumentation), more focus shifts to non-task related issues, and they were less constructive than the asynchronous discussions in Allaire Forums. The asynchronous discussions were only ‘indirectly’ argumentative (including information checks), they maintained a more conceptually oriented focus and contained more constructive activities. To maintain a conceptually oriented focus and to co-ordinate interactions appears to be particularly related to the asynchronous and synchronous modes of communication. In synchronous discussions students engage in a fast flow of communication. Real-time pressures them (psychologically) to read and respond to each other’s contributions within seconds or at most minutes. Focus shifts to non-task related aspects or technical issues easily cause students to lose track of an argument or to lose the overview of the main issues under discussion. In asynchronous discussions students may take hours, days, weeks, and sometimes even longer to read, write and think about contributions that triggered their interest, instead of seconds or minutes. More time may afford re-reading and reflection, keeping track of the line of discussion and treating non-task related interactions or technical disturbances for what they are: temporary, peripheral interruptions.
3.2.4 Support in CMC Systems
In all three studies human or interface support primarily aimed at promoting argumentative processes. However, no effects were found from human ‘structure’ coaches who supported the taking of multiple perspectives and counter-argumentation. ‘Reflective’ support increased the number of check questions asked, which later turned out to be powerful in relationship to the production of constructive activities. Graphical support on the Belvédère interface triggered students to produce more counter-arguments, a ‘direct’ form of argumentation. However, counter-argumentation was not related to effective student discussions. Then again, the Belvédère discussions were relatively more often conceptually oriented and constructive than the NetMeeting discussions. Perhaps the separate window for argumentative diagram construction particularly facilitated focus maintenance, and subsequently stimulated the production of constructive activities. It may be possible, however, that a tool for regular concept mapping1 might have been just as effective as the diagram construction tool (Van Boxtel, 2000). It is not known (yet) to what extent the beneficiary effect is due to particular constructs of the Belvédère system.
To refer back to one of the earlier points mentioned, relatively speaking the asynchronous discussions in Allaire Forums were more often conceptually oriented and constructive than the synchronous Belvédère and NetMeeting discussions. The NetMeeting discussions were most often focused on finishing the task. However, in clustering the discussions on the factors of focusing, argumentation and constructive activities, some discussions in Allaire Forums were also found to be less effective; some Belvédère discussions were completely product-oriented and a few NetMeeting discussions were even found to be highly conceptually oriented and constructive. This indicates that in addition to features of the electronic systems and task characteristics, effective discussions also relate to individual group differences, such as task approaches, preparation activities or collaboration strategies, and to factors of the broader educational context.