Functional versus spontaneous roles during CSCL

In this paper, two studies are reported on the effect of functional roles on computersupported collaborative learning (CSCL) in higher education – the second is a replication of the first. Prescribed functional roles were implemented in half of all groups during a project-based course in higher education. All communication was via e-mail. Analysis of Likert-scale evaluation questionnaires gathered in both studies revealed a latent variable ‘perceived group efficiency’ (PGE) which – depending on the level of constraints set by preconditions – appears to increase the awareness (Study 1) or the level of efficiency (Study 2). However, Likert-scales provide a surface level analysis of actual behaviour and no insight in the collaborative process. Hence, the e-mail communication was investigated with two content analysis procedures: content analysis of the type of communicative statements and analysis of the role behaviours performed in role and nonrole groups. Results from both studies reveal that significantly more statements are focused on coordination in role groups. In addition, analysis of role behaviour reveals that students in role groups perform significantly more according to the functional roles than their counterparts in nonrole groups, although spontaneous role behaviour emerged in nonrole groups as well.