A. Figueira

University of Porto (PORTUGAL)
Group work is a fundamental activity for the development of organizational and criticism skills. From kindergarten up until university level, group work is generally employed by educators in this whole spectrum. It is also accepted that the development of “social skills” is fostered by participating in groups, and by collaborating with others in some joint work.

In the last two decades research on online group work has been carried out and has systematically growth in importance. Accordingly, there is a vast theoretical literature, numerous case studies, and many tools that handle (online) and help manage group work.

However, although tools and analysis of the tools to manage group work have largely been proposed, we haven’t yet seen much concern with the assessment of group work, particularly of individual assessment of group participants. We may say that more than frequently, the assessment of the group is transposed to each group participant, which in turn results in each student having the same final mark.

This situation poses many problems and eventually some injustices: there may be many students that do not contribute with good work, or even that do not contribute at all. On the other hand, there may be students that try to participate actively but whose opinion was abrogated by the majority of the group. We recall that it is not always easy to have the perception of the contribution of each student to the whole group work. However, the merits of this kind of pedagogical activity largely justifies the need design reliable instruments and techniques to provide a greater accuracy (and legitimacy) when assigning a single student’s final mark.

In these settings and context we propose a novel tool to manage and assess group, taking into account the amount of work, the group interaction, the quality and the temporal evolution of the participation of each group participant, and also profile comparison with other online collaborative tasks.

Our developed tool features the possibility to create collaborative or cooperative work (it is up to the teacher to decide which suits best, at any case), both in the Moodle LMS. The collaborative work mode is based on the vision that the work is carried out by applying systematic transformations from an initial state by all group members, up to a state commonly accepted by the group as “final”. On the other hand, the cooperative work is based on the perspective that the objective can be divided in a guided “pipeline” of tasks which are then dynamically distributed among the group participants. All these actions are supervised automatically by the tool through an a priori configuration.

In the article we describe and discuss the philosophy of our tool, which is based on events, notifications and on pattern mining from the log data. From this basis we describe the concept of the assessment methodology, which operates on the combination of teacher and peer evaluation. We then introduce the data mining techniques which are employed by the tool to detect student interaction patterns, and compare them with stored patterns of previously assessed group work in order to detect abnormal deviations.

We conclude the article presenting the tool’s strengths’ and weaknesses’ from a more philosophical perspective, and point out directions for future developments.