Abstract
This article explains and implements a network analytic approach to the study of cross-platform audience behavior. It begins by conceptualizing large-scale patterns of media use in network terms, treating media outlets as nodes and the levels of audience duplication among them as links. Following that, it explains 2 common measures of audience duplication, Absolute Duplication and Primary Duplication, and offers a new measure, Deviation-from-Random Duplication. In doing so, techniques for converting duplication data into network data are discussed. This approach is then applied to analyze patterns of audience fragmentation, media publics, and audience polarization using data from Nielsen's TV/Internet Convergence Panel. The findings show the value of using a network approach, by contributing to an alternative understanding of these patterns. Economic and policy implications are discussed, as well as broader reflections on the use of network analysis in the study of audience behavior.
Notes
1 Disclosure: Access to this otherwise proprietary data is the result of a data use agreement with the Nielsen Company. The agreement gives me full intellectual property rights, although Nielsen reviews the written work to ensure it represents their panel and collection procedures accurately. The findings are unhindered by this process.
2At this point, Nielsen does not track Internet use on the panelists' work computers. In addition, the tracking software is currently only available for PC users. A Mac version is in development. Primary computer is defined as the PC that is used most in the household.