Perceptions of and Exposure to Misinformation

In this study, we identify the determinants of individuals’ perceived exposure to misinformation on social media. While misinformation is not a new phenomenon, social media has proved particularly effective at enabling its spread globally. Several studies and surveys show that a large share of social media users across several countries are concerned about the spread of, and believe that they have been exposed to, misinformation online. Yet, studies based on browsing and digital trace data conclude that the problem of misinformation on social media is overstated or has decreased (Acerbi et al. 2022; Allcott et al. 2019; Allen et al. 2020; Allen et al. 2021; Eady et al. 2023; Grinberg et al. 2019; Guess et al. 2019; Guess et al. 2020; Moore et al. 2023). These contradictions could come from the fact that misinformation tends to spread faster and receive high levels of engagement (Edelson et al. 2021; Marchal et al. 2019; Vosoughi et al. 2018), that audio-visual content is rarely accounted for in the above studies (see DiResta et al. 2019; Garimella and Eckles 2020; Nimmo et al. 2020; Reis et al. 2020; Yang et al. 2023 on audio-visual misinformation), or that misinformation topics are treated separately (e.g., some studies on political misinformation dismiss politically-charged topics, such as medical misinformation; see Gallotti et al. 2020; Kouzy et al. 2020; Li et al. 2020; Loveland et al. 2024; Madraki et al. 2021; Nsoesie et al. 2020; Singh et al. 2022). With these findings in mind, in this study we investigate individuals’ perceived exposure to misinformation while controlling for their actual exposure to it, which we approximate in several ways. We do so by combining survey data with browsing data and social media takeout data. We argue that perceptions of being exposed to misinformation may be heightened when individuals are exposed to information from their opposing ideology. We build on the literature on ideologically-motivated reasoning which argues that individuals have a hard time identifying misinformation that aligns with their preexisting beliefs and political views, and tend to discredit information that contradicts their views (Gampa et al. 2019; Gawronski 2021; Kunda 1990; Taber and Lodge 2006; Taber et al. 2009). We conduct an individual-level survey in Spain (N = 1,700), and combine the survey data with browsing data and social media takeout data donated by a subset of participants. This allows us to approximate individuals’ actual exposure to misinformation or low-quality news from different sources online. By combining multiple data sources, we make a large contribution to the field by providing and measuring the effect of alternative measures of misinformation exposure.