Patterns in Engagement with Verified False Content on Facebook across the EU

The next session at AoIR 2023 is our own panel, and starts with a presentation by Jessica Walter and Anja Bechmann. Their focus is on influence processes surrounding verified false content across the EU, with particular focus on national differences between EU countries as well as differences driven by other demographic factors. The EU is relatively understudied with respect to the influence of mis- and disinformation, compared to the US and other countries.

The distribution of verified false content represents a case study of unwanted influence; the present study focusses on false content on Facebook as identified by Meta’s third-party fact-checking network, and draws on the URL shares dataset made available by Facebook via Social Science One; this covers the period of 2017-22, and contains links that were shared at least 100 times publicly (which will privilege larger EU countries that will meet that threshold more easily). The dataset contains some basic demographic information about the content sharers as well.

Top countries showing up in the dataset are France, Greece, Germany, Spain, and Poland, which saw between 1700 and 500 unique verified false content URLs per country being shared. Men account for a higher percentage of VFC exposure than non-VFC exposure, although overall women are viewing the greater amount of VFC content; this might be important for developing interventions to combat mis- and disinformation. VFC is also being shared by somewhat older audiences than non-VFC content. There are also considerable differences between the types of content viewed by different age groups – younger users view more sensationalist VFC, while older users view more political and clickbait VFC.

There are also substantial differences between European regions: Nordic countries are quite US-oriented in their VFC viewing; Eastern Europe focusses more on Russia/Ukraine VFC; central European countries focus more on refugee issues; and Mediterranean countries focus more on health issues. Such exposure also varies over time, of course, with particular changes observable during the COVID-19 pandemic and with the start of the Russian full-scale invasion of Ukraine.

Further finer-grained analysis is still necessary to explore more detailed patterns in VFC exposure as revealed by this dataset; however, the dataset itself is also limited in its detail, and further additional data gathering may be needed.