Collaborative Writing

Analysing a collaborative writing activity in order to improve tutor’s perception of individual contributions of learners Christelle Laperrousaz, Pascal Leroux & Philippe Teutsch Laboratoire d'Informatique de l'Université du Maine - CNRS FRE 2730 University of Le Mans, France {Christelle.Laperrousaz,Pascal.Leroux, Philippe.Teutsch}@lium.univ-lemans.fr

Abstract This paper is interesting in the tutor’s perception of the learner’s contribution to a collective work. In particular, we propose means of individual perception of a collaborative writing activity. We present what a collaborative writing activity is and what the tutor’s needs of perception are in order to follow a collaborative writing activity. Then, we present our propositions of perception of individual contribution through the analysis of discussions.

1.

Introduction

During the follow-up of a distance collective activity, the tutor needs to understand directly the learner’s individual progression, in order to support each learner in the collective task. The success or the failure of a collective work is evaluated from the contribution of individuals to the elaboration of the common task [1]. To evaluate the learner’s contribution, the tutor has to recognize each learner in the final collaborative production. For that, s/he needs some tools which help him/her perceive each learner’s activity as clearly as possible. We propose that the tutor’s perception of the learner’s activity be possible through the consultation of the learner’s individual productions, the consultation of his/her contribution to the discussions and collective productions and through the visualization of his/her place in the group dynamics. Our context of work is the collaborative writing in French as a Foreign Language field. So, we take an interest in describing how to perceive individual contribution to a collaborative writing activity. First, we describe what a collaborative writing activity is, and what the tutor’s needs to perceive each learner’s contributions are. Then, we present our propositions to analyse the learner’s contributions.

2. 2.1.

Collaborative writing activity Common tasks of collaborative writing

T. Cerratto [1] proposes two main tasks of collaborative writing: the planning of collaboration and the production of the collaborative document. The planning of collaboration is composed of discussions about the agreement of the goal, the tasks’ division (choice of the writing strategy, and assignment of roles to group members), the planning of synchronous or asynchronous meetings. Producing the collaborative document involves making the plan of the document (proposition and structuring of individual and group ideas), writing the document and reviewing it. So, exchanges between group members concern planning, writing and checking of the content.

2.2.

Writing strategies

Four writing strategies can be envisaged [2]. Group single-author writing strategy occurs when one member is directed to write for the group: this strategy is used when consensus on the written document is not very important and when the writing task is simple. Sequential single writing occurs when one member writes at a given time: each writer completes his/her part of document and then passes it to the next writer. This strategy simplifies the organization but can create a lack of consensus and implies frequent controls of the content. Parallel writing occurs when the production is divided into different parts and when members write in parallel: each participant is responsible for a part. Reactive writing occurs when members write the document together, reacting and commenting to each other. The strategy can be imposed by the tutor or by the collaborative tool, or chosen by the group.

Proceedings of the Fifth IEEE International Conference on Advanced Learning Technologies (ICALT’05) 0-7695-2338-2/05 $20.00 © 2005 IEEE

2.3.

Roles of learners

Learners can play various roles during the collaborative writing. The writer [2] is responsible of the translating of ideas from discussion to a part of the collaborative document. The consultant participates actively during the writing stages but has no responsibility for the content production. The editor is responsible for the overall content production; s/he provides corrections to document parts of other group members. The reviewer comments the collaborative document, on both the content and the form: the writer accepts or not these comments. The group leader fully participates in authorship and reviewing activity, but also leads the group, responsible of the planning management, and encouraging learners’ motivation. The facilitator is external to the collaborative writing; s/he does not give content feedback: for us, this person is the tutor of the group. Roles can change during the collective activity. They can be proposed by the tutor, or distributed by the group among group members. In each case, it can be interesting to provide the tutor with information about the role of each learner, exploring the discussion. With such information, the tutor would be able to check if a learner plays the role s/he has assigned to him/her. Moreover, s/he would be able to encourage a learner to play a specific role.

2.4.

Tutor’s needs and tools

In order to perceive the learner’s contribution in the group work and to evaluate it, the tutor needs to understand the collaborative writing process and the learner’s place in it. In particular, s/he needs to know the role(s) played by the learner during discussions and productions. The tutor’s needs are dependant on the roles s/he plays to follow the activity, on the writing strategy, on the stage of the collaborative process, and on the collaborative writing environment used by the group. S/he needs information about the group coordination (writing strategy, role distribution, planning), and about the group production (plan of the document, individual and collective ideas, comments). Currently, the tutor does not have a lot of analysis tools to facilitate him/her in his/her task of perception. The literature presents works about quantitative and qualitative analysis of the contributions made by the learners during the collective production. Works of [3] present a tool of analysis of the group. First, a quantitative analysis concerns the learner’s contributions during a discussion (using a sentenceopener discussion tool): evolution of the number of contributions, contribution size and number of contributions answered by other members. Second, a

qualitative analysis concerns the analysis of individual learner’s behaviours compared to the rest of the group members. Other works combine quantitative analysis of interactions with learners’ impressions about the collaborative process and the collaborative production [4] to better understand the collaborative process. We propose to analyze discussions from an individual point of view, categorizing the learner’s interventions and identifying the learner’s roles. We present our proposals in the next section.

3.

Analysis of the learner’s contributions

From the literature, we have noted that most of the collaborative writing environments allow the group to adopt a parallel writing strategy [4] [5] [6], so, in the rest of the paper, we restrict our study to collaborative writing activity whose writing strategy is parallel. In such a writing strategy, most part of the collective work is made during discussions: analysing the learner’s contributions implies analysing discussions from an individual point of view. Discussions can be seen as a shared understanding of the task to be carried out and of the current activity. During collaborative writing, exchanges between co-authors are both means and products of the collaboration in the document writing [7]. Each individual contribution is subjected to the judgement of others that arouses approval, reject or a counterproposal [8]. The tutor has also to take this judgement into account in order to evaluate the learner’s contributions. We think that exchanges constitute a relevant trace to understand the collaborative writing process, and we propose to analyze both the learner’s interventions and the reactions of others. To facilitate the identification of communicative acts, we base our analysis on tools of discussion semi-structured by communicative acts.

3.1.

Analysis of the learner’s interventions

One of the difficult things but interesting for the tutor is to characterize learners’ interventions (relating to the writing task or relating to the interaction for example). Cognitive sciences propose manual categorizations of interventions in discussion. T. Cerratto [1] identifies three categories: interventions about action, about the content and about the artefact. Interventions about action are composed of initiative interventions about the way of writing the collaborative document. Interventions about the content are composed of initiative interventions about what to write. Interventions about the artefact are composed of initiative interventions about functionalities of the

Proceedings of the Fifth IEEE International Conference on Advanced Learning Technologies (ICALT’05) 0-7695-2338-2/05 $20.00 © 2005 IEEE

collaborative environment. M. Quignard [8] envisages two types of interventions: epistemic interventions which introduce new knowledge elements, axiological interventions which express a judgement or a problem evaluation. In particular, he proposes a collaborative model of argumentation, analyzing an intervention through three criteria: its universe of reference (the problem-solving task versus the dialogue structure), its critical thinking operation (an evaluation versus a making explicit), and its interlocutory orientation. At any time, a learner can contribute to the main problem or to the dialogue control. We think that it is interesting to provide the tutor with an automatic classification of the learner’s exchanges in order to help him read the discussion. We propose to distinguish interventions relating to the writing task and interventions relating to interaction. In each category, we propose to identify propositions, evaluations, and explanations. Moreover, we propose to characterize exchanges as minimal or complete. A minimal exchange is composed of an initiative intervention and a reactive intervention. A complete exchange is composed of an initiative intervention, a reactive intervention and an evaluative intervention. Complete exchanges characterize a more important undertaking of the learner who initiates them [6]. So, distinguishing minimal and complete exchange is a means for the tutor to evaluate the learner’s undertaking.

3.2.

Analysis of the learner’s roles

Another type of useful information for the tutor is the learner’s roles, analysing discussions. To identify the learner’s role, we propose to base our work on the ethology work of Pléty [9] who has analyzed interactions of schoolchildren working in groups of four to solve algebraic problems and has highlighted four social behaviour profiles according to the learner’s volume of intervention, to the learner’s type of intervention, to the learner’s communicative gestures and to the reactions of the other learners. The four social behaviours are the Moderator, the Seeker, the Valuator and the Independent. From this ethology work, a first computer study has been done by George and Leroux [10] who have developed a tool which makes it possible to perceive learners’ social behaviours during collective discussions semistructured by communicative acts. Analysis is graphically represented. In order to calculate the behaviour profiles, the tool we propose glances through the collective discussion and for each intervention, keeps in memory the speaker’s name, the communicative act, and the reaction of others to this intervention. For each reaction, the tool keeps in

memory the speaker’s name and the communicative act. Depending on the communicative act used, the tool enriches one of the behaviour profiles of the learner who has intervened. The tutor has various kinds of views at his/her disposal: visualization of the behaviour profile rates during one collective discussion, visualization of the evolution of the rate of behaviour profiles as the collective discussions go on.

4.

Conclusion and future work

We have presented what a collaborative writing activity is and how to improve the tutor’s perception of each learner’s contribution. We have proposed several analyses of discussions in order to highlight individual contribution to a collaborative writing document, especially when the writing strategy is parallel. Currently, we are developing an environment which will integrate all our propositions of perception.

5.

References

[1] Cerratto, T. (1999). “Activité Collaborative sur Réseau, Une approche instrumentale de l'écriture en collaboration”. Thèse de doctorat, Université Paris VIII, France. [2] Lowry, P. B., Curtis, A., Lowry, M. R, (2004), “Building a taxonomy and nomenclature of collaborative writing to improve interdisciplinary research and practice”, Journal of Business Communication, 41(1), pp. 66-99. [3] Barros, B., Verdejo, M. F. (2000), “Analysing student interaction processes in order to improve collaboration, The DEGREE approach”, International Journal of Artificial Intelligence in Education, 11, pp. 221-241. [4] Santoro, F. M., Borges, M. R. S., Santos, N. (2004), “Evaluation of collaborative learning processes”, Advanced Technology for learning, 1(3), pp. 164-173. [5] Hofte, G. H., Van Der Lugt, H. J. (1997). “CoCoDoc: a framework for collaborative compound document editing based on OpenDoc and CORBA”, IFIP/IEE International Conference on Open Distributed Processing and Distributed Platforms, Toronto, Canada, pp. 15-33. [6] Rodriguez, H., Severinson-Eklundh, K. (2002), “Supporting individual views and mutual awareness in a collaborative writing task: the case of Col-laboració”, Technical report TRITA-NA-P0214, NADA. [7] Sharples, M. (1993), “Introduction”, In Computer Supported Collaborative Writing (Springer-Verlag), pp. 1-7. [8] Quignard, M. (2002), “A Collaborative Model of Argumentation in Dyadic Problem-Solving Interactions”, ISSA'02, Amsterdam, The Netherlands. [9] Pléty, R. (1996). L'apprentissage coopérant, Presses Universitaires de Lyon ed., Lyon, France. [10] George, S., Leroux, P. (2002), “An approach to automatic analysis of learners' social behavior during computer-mediated synchronous conversations”, ITS'02, Biarritz (France), pp. 630-640.

Proceedings of the Fifth IEEE International Conference on Advanced Learning Technologies (ICALT’05) 0-7695-2338-2/05 $20.00 © 2005 IEEE