Personalized News: How Filters Shape Online News Reading Behavior.

摘要

The evolution and diffusion of communication technology has consistently changed interactions between members of the public sphere in forming public opinion. Some democratic scholars have worried recent developments in personalization technologies will degrade public opinion formation. They worry that personalized news allows citizens to only pay attention to news coming from their preferred political perspective and may isolate them from challenging perspectives. Empirical research has shown people with access to more highly selective information technology demonstrate increases in both selectivity and incidental exposure to diverse perspectives.;This dissertation focuses on these behavioral and attitudinal outcomes of using personalized news technologies. Dual-processing theories of information provide the foundation for analyzing opinion formation within the bounded rationality model of public opinion. Personalized news technologies are hypothesized to increase the amount of news exposure and elaboration through increased personal relevance.;Two studies test these broad hypotheses. First, results from a national random sample of adults show users of personalized web portals are more likely to engage in increased news viewing both online and offline. No differences in preference for perspective sharing or challenging sources of news is found between personalized portal users and non-users. Next, results from an online experiment of Ohio adult Internet users show an increase in time spent reading news articles in personalized news portals compared with a generic portal. An interaction between using customized news portals with source recommendations based off of explicit user preferences and increased time spent reading per news article is found on news elaboration. No differences in news elaboration are found in other personalized news designs including implicitly recommended news sources based on user profile information and only showing users recommended stories. The implications of these results are discussed in terms of the public opinion debate about new communication technologies, selective exposure research, information processing research, and personalized information system design.