October 2008 - December 2012
Leibniz Graduate School for Knowledge Media Research, budget resources KMRC
Informed decision making and elaborated opinion formation on controversial issues require knowledge about the diversity of opinions available. The Web is a perfect backdrop for opinion formation as a multitude of different opinions is publicly available, i.e. in discussion forums and in social networks. However, the different opinions often remain unexploited When people seek new information, these information search processes are often biased in favor of the information seeker’s previously held expectations; an effect that has been labeled confirmation bias.
This project is based on the assumption that the use of technologies like recommender systems are a promising approach to overcome this bias and to stimulate elaboration. Of prior importance is the role of preference-inconsistent recommendations compared to preference-consistent recommendations. In previous studies it has been shown that preference-inconsistent recommendations can reduce confirmation bias. Further, preference-inconsistent recommendations affect preferences and elaboration, i.e. they lead to more divergent thinking. Moreover, barriers and facilitators for the impact of preference-inconsistent recommendations were investigated. Results show that preference-inconsistent recommendations are especially effective for learners with no prior knowledge or under cooperation.
Based on the results, implications for developing recommender systems should be derived.