The theorizing about digital media behavior and information consumption depends greatly on how information is collected and classified. Much of our understanding of digital media behavior has been based on aggregate data that is platform-specific and mostly decontextualized, studying behavior from within the boundaries of the media product rather than the experience of media consumption. Using factual vs. non-factual information as a characterization of smartphone use data, this study examines digital media behavior across platforms and in situ. Based on a dataset of 1.5 million smartphone screenshots collected from 94 American adults, the consumption of factual information is explored in its association with three types of variables: demographics, time of use, and online media platforms. Findings of this study are assessed and discussed in the context of the theories of media processing and information consumption.

Keywords: Media Behavior, Smartphones, Computational Methods, Machine learning.