Exposure to political diversity online.

다국어 초록 (Multilingual Abstract)

The Internet gives individuals more choice in political news and information sources and more tools to filter out disagreeable information. Citing the preference described by selective exposure theory—that people see and attend to information that supports their beliefs and avoid counter-attitudinal information—observers warn that people may use these tools to access agreeable information and live in ideological echo chambers, increasing the polarization of different political groups and decreasing society's ability to solve problems.
This dissertation studies political information exposure in two types of online spaces. First, it examines online news aggregators, where people's political preferences will shape their exposure. It describes individuals' preferences for the range of political opinions news aggregators present, ways to measure the diversity of exposure in those spaces, and selection and presentation techniques for increasing the diversity of exposure. Second, it discusses non-political spaces, where preferences other than politics shape people's behavior, but where people may still serendipitously encounter political information.
This work contributes to both understanding people and building systems. First, addressing mixed results within the selective exposure literature, it finds that people are neither inherently challenge-averse nor inherently diversity seeking; there are individual differences. In non-political spaces, it also finds substantial political discussion on non-political blogs, where people may have serendipitous encounters with diverse views. Moreover, blog readers do not treat these posts as taboo and they engage with the posts' political content. This argues that serendipitous encounters with mixed viewpoints will still happen, even if not in news aggregators: even if efforts to intervene and increase the diversity of exposure on news websites fail, scholars should not be so alarmed.
For designers and builders of online political news tools, and other applications, the dissertation proposes diversity metrics including inclusion, alienation, and representation scores. It also describes and presents an evaluation of the Sidelines algorithm, designed to select diverse collections from user votes. Finally, it describes a design space for visualizations intended to nudge people to read more balanced or diverse sets of news, an evaluation of two such techniques, and a research design for field evaluation of further visualization techniques.