Source memory and collaborative learning: The role of group composition and conflicting information
Tue-HS3-Talk V-02
Presented by: Oktay Ülker
Source memory is important in collaborative learning scenarios: when learners receive new information from learning partners with differing knowledge-levels, remembering the source helps to better assess the reliability of information in retrospect. Moreover, learners sometimes receive information from partners which is conflicting with prior received information. This enhances source memory, as receiving conflicting information guides attention to sources. In a 2×2 between-subject experiment, we examined the effects of group composition (three partners with differing knowledge-levels vs. same knowledge-levels) and conflicting information (with conflict vs. without conflict) on source memory and knowledge acquisition in a pseudo-collaborative learning scenario (N = 128). Three bogus partners presented texts with information which either contradicted information participants received earlier or did not. Group composition and conflicting information did not influence knowledge acquisition, but both factors influenced source memory: analyzes with multinomial processing tree models revealed that participants better remembered which partner presented certain information in groups with differing knowledge-levels. High-knowledge and low-knowledge partners were remembered better than medium-knowledge partners. However, only the source memory advantage for high-knowledge partners was significant, partially suggesting that in collaborative learning, remembering which information might be right is rendered more useful than remembering which information might be wrong. Contrary to our hypothesis, source memory was worse in a context with conflicting information. Our findings contribute to a better understanding of context-dependent source memory and thus, consequently, to the question how remembering whether certain information might be right or wrong is adaptively prioritized in different contexts.
Keywords: Collaborative learning, Multinomial modeling, Source memory