The blind spots of measuring online news exposure: a comparison of self-reported and observational data in nine countries

ABSTRACT

Measures of news exposure are common in research that tries to explain political knowledge, political engagement, opinion formation and, more generally, media effects. Much of that research employs self-reported measures obtained with surveys, known to suffer from accuracy problems. Observational measures, however, also suffer from limitations derived from data collection and instrumentation. Here we offer new comparative evidence on the nature of those problems. We show that commonly used self-report measures of digital news consumption are problematic for three reasons: they only pay attention to a small fraction of all available sources; they underestimate audience share; and they distort the relative position of news sites in visibility rankings. Measurement problems, however, also exist in observational studies, especially when mobile access is excluded from data collection. Our analyses quantify the magnitude of these problems, offering unprecedented comparative evidence of online news consumption that spans nine countries and a period of five years. We discuss the implications of our findings for future research on news exposure.

KEYWORDS:

• News consumption
• web
• digital traces
• exposure
• measurement error

Acknowledgements

The idea for this paper originated while the authors were visiting the Weizenbaum Institute in Berlin, Germany in the summer of 2019. We would like to thank Martin Emmer, Annett Heft, Curd Knüpfer and Pablo Porten-Cheé for their invitation to visit.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work has partly benefited from funding from the National Science Foundation (grant numbers: 1729412 and 2017655).

Notes on contributors

Sandra González-Bailón

Sandra González-Bailón is an Associate Professor at the Annenberg School for Communication, University of Pennsylvania. Her research and teaching interests lie in the area of political communication and computational social science.

Michael A. Xenos

Michael Xenos is a Professor in Life Sciences Communication at the University of Wisconsin-Madison. His research and teaching interests are centered on the effects of digital media on political engagement and public deliberation.