Collaborative Learning is believed to have positive effects, due to (among other things) processes of social co-construction of knowledge such as externalization and elaboration, elicitation, conflict and integration (c.f. Fischer et al., 1998; Weinberger & Fischer, 2006). However, negative effects of collaborative learning and computer supported collaborative learning (CSCL) are also documented (Renkl et al., 1996; Hron et al., 2002). In order to support cooperative learning, Slavin (1992) assigns group rewards on the basis of individual responsibility which obtains its positive effect through motivation.

There is convincing empirical evidence that motivation influences learning outcomes in the context of conventional learning scenarios (c.f. Schiefele, 1996). The cognitive-motivational model of learning (Rheinberg et al., 2000) postulates that enduring motivational attitudes influence actual motivation. Actual motivation in turn affects knowledge acquisition via the mediator's time on task, learning activities and functional state.

In this study, the model by Rheinberg et al. (2000) was applied to collaborative learning by identifying the „learning activities“ of the original model with processes of social co-construction of knowledge. In order to examine the relationship between motivation and knowledge acquisition, actual motivation was manipulated by choosing a learning topic which should produce a great variability in learners’ interest and by experimentally varying the presence of a group reward sense Slavin (1992). As the dependent variable and process variables respectively, knowledge acquisition, time on task, functional state and processes of social co-construction of knowledge were measured. A positive interrelation between on-line motivation and knowledge acquisition during CSCL was found. However, group rewards as well as the analyses of processes of knowledge co-construction showed unexpected effects. In order to incorporate these results, the underlying model is modified and discussed by taking into account both individual and collaborative learning phases and processes.

This research refers to a rarely examined combination of two domains of educational research: (Computer Supported) Collaborative Learning and Motivation. Collaborative Learning is believed to have positive effects, due to (among other things) processes of social co-construction of knowledge such as externalization and elaboration, elicitation, conflict and integration (c.f. Fischer et al., 1998; Weinberger & Fischer, 2006). However, negative effects of collaborative learning (such as the free-rider-effect) are also documented (Renkl et al., 1996). Computer-mediated communication leads to additional problems, such as lacking social presence (c.f. Hron et al., 2002). In order to support collaborative and cooperative learning, various approaches were developed. One of those is that of Slavin (1992) who assigns group rewards on the basis of individual responsibility. His approach follows motivationally oriented concepts of support and has positive effects on cooperative learning (Slavin, 1996).

Following Slavin’s (1992) proposal that motivation is key in the process of supporting collaborative learning is the second research domain: the influence of motivation on learning. There is convincing empirical evidence that motivation influences learning outcomes in the context of conventional individual learning scenarios (c.f. Schiefele, 1996). Furthermore, mediating this interrelationship between (intrinsic) learning motivation and knowledge performance is probably the application of learning strategies (c.f. Schiefele & Schreyer, 1994). The cognitive-motivational model of learning (Rheinberg et al., 2000) postulates that enduring motivational attitudes influence actual motivation. Actual motivation in turn affects knowledge acquisition via the mediators time on task, learning activities and functional state.

Research Question. This theoretical background concerning motivation on the one hand and collaborative learning on the other hand leads to the main research question: How does actual motivation influence knowledge acquisition and collaborative processes of knowledge acquisition (processes of social co-construction of knowledge) during computer supported collaborative learning (CSCL)? As a research model, the model by Rheinberg et al. (2000) was applied to collaborative learning by identifying the „learning activities“ of the original model with processes of social co-construction of knowledge. Following this model, actual motivation should increase knowledge acquisition. This relationship should be mediated by time on task, functional state and social processes of co-construction of knowledge.

Study. In order to examine the relationship between motivation and knowledge acquisition, actual motivation was manipulated by choosing a learning topic which should produce a great variability in learners’ interest and by experimentally varying the presence of a group reward sensu Slavin (1992). As the dependent variable, knowledge acquisition was measured by pre- and posttest, process variables obtained were time on task, functional state and processes of social co-construction of knowledge, measured by analysing logfiles following the coding scheme of Weinberger and Fischer (2006). It was hypothesized that group rewards should increase actual motivation. Moreover, actual motivation – which should also be influenced by the learner’s interest in the learning topic – should lead to higher knowledge acquisition as well as longer time on task, higher scores of flow (indicating functional state) and better collaboration indicated by processes such as explanations or integration of different points of view.

40 university students took part in the study. Their task was to develop in pairs a handout about a statistical topic. The experimental group were additionally provided with a financial group reward on the basis of their post knowledge tests and the quality of their handout. After completing some pre-questionnaires, participants received a short introduction into the software tools and the instruction for the task. They then filled in the FAM (questionnaire of actual motivation, Rheinberg et al. , 2001). Following this, participants were presented with short introductory texts (different for each subject of a pair in order to generate „distributed knowledge“, c.f. Reinmann-Rothmeier & Mandl, 2002) and had to collaboratively develop a handout with the structure and main definitions of the topic by only communicating via computer. Only a chat tool and an editor on our course management system (Moodle) were provided. During the task, participants had to fill in a short questionaire on their on-line motivation and their functional state (flow). After finishing the task, participants completed the post knowledge test and filled in some questionnaires about their learning activities during the task.

Results. As expected, a positive correlation between on-line motivation and knowledge acquisition during CSCL was found. By contrast, group rewards did not show the expected effects (in a multivariate ANOVA). Instead of increasing actual motivation and knowledge acquisition, group rewards led to lower perceived probability of success (measured by a scale of the questionnaire of actual motivation). Moreover, there was no influence of group rewards on knowledge acquisition. However, detailed analyses revealed that the group reward participants worked harder on the handout.

Findings did not support the expected role of time on task and functional state as mediators of the relationship between motivation and knowledge acquisition. Furthermore, the interrelationship of motivation and knowledge acquisition on the one hand with processes of social co-construction of knowledge on the other hand turned out to show unexpected effects: Findings show no or even negative effects of actual motivation on desirable processes of co-construction of knowledge (as elaborations, explanations, elicitations, detection and integration of different points of view,…). For example, the more motivated people were, the less questions they asked.

Discussion. The unexpected effects of group rewards could be explained in two ways: Maybe participants felt more under pressure in the group reward condition, so that their perceived probability of success decreased and they worked harder on the handout, but with no further remarkable effect on knowledge acquisition or other learning processes. Another explanation could be that giving them a concrete goal (75% of points) led to more realistic self appraisals.

At first glance, the cognitive-motivational model of learning seems to be unconfirmed for cooperative learning as no mediators were found. However, these contradictory results reveal the possible influence of uncontrolled individual phases of learning. In order to incorporate this assumption, the postulated model is modified and discussed by taking into account both individual and collaborative learning phases and processes.