Designing for diversity in online news recommenders

Information diversity has a long tradition in human history. Recently there have been claims that diversity is diminishing in information available in social networks. On the other hand, some studies suggest that diversity is actually quite high in social networks such as Twitter. However these studies only focus on the concept of source diversity. Further, personalization algorithms such as online news recommenders, if designed incorrectly, can decrease information diversity. In this paper, we study different dimensions of the value information diversity. Later, using Value Sensitive Design methodology, we translate this value into norms, which can be used as design requirements.