Prof. Magdalena Wojcieszak from the Centre of Excellence in Social Sciences at the UW and the University of California, Davis, together with an international team of researchers, has conducted research into the exposure to and impact of content from untrustworthy sources on social media users. The results of the analysis were published in the journal “Science Advances”.

Prof. Magdalena Wojcieszak from the Centre of Excellence in Social Sciences at the University of Warsaw and the University of California, Davis, is a co-author of the article Untrustworthy sources on Facebook and Instagram in 2020: Concentrated exposure but no attitudinal effects, which appeared in the journal “Science Advances”.

An international team of researchers conducted a study into content published on social media and its impact on active adult users of Facebook and Instagram in the United States in the run-up to, during and immediately after the 2020 presidential election. The study consisted of two parts: a data analysis and a field experiment involving active adult users in the United States who use Facebook and Instagram. The researchers wanted to ascertain the extent of exposure to disinformation.

The researchers point out in their publication that the average user saw relatively little content from unreliable sources in their social media feeds. The field experiment, which lasted several months and was conducted amongst users who had consented to take part in the study during the 2020 presidential election period, reduced the visibility of content from unreliable sources in the news feeds on both platforms by around 70 per cent. However, it did not yield measurable results in relation to a number of pre-defined indicators – even among study participants who had frequent exposure to such content prior to the experiment.

The study’s findings show that measures which can be implemented by platforms can effectively limit exposure to content from unreliable sources, although these changes are unlikely to have an immediate impact on attitudes and beliefs.

“Both this article and others we have published as part of The U.S. 2020 Facebook & Instagram Election Study are the result of a unique collaboration with Facebook (later Meta). Thanks to this collaboration, we gained access to proprietary data from the platforms, which is rarely available to researchers. All our research shows that changes to social media platforms’ algorithms have a huge impact on what people see and do in their news feeds, but do not influence users’ political opinions and attitudes. This specific study, described in ‘Science Advances’, further demonstrates that concerns about so-called ‘echo chambers’ or ‘misinformation filter bubbles’ are not entirely accurate. Most people are simply not very interested in politics, consume little political information and do not find themselves trapped in ‘information bubbles’ dominated by disinformation. People who encounter such content already hold strong, well-formed views, which may motivate them to deliberately seek out material that reinforces their existing beliefs,” says Prof. Magdalena Wojcieszak.