When do fact-checks work to counter misinformation? A cross-national experiment on the influence of the source and degree of politicization of fact-checks

While not being a new phenomenon, the past years have seen a strong surge of mis- and disinformation. One solution that is often suggested to counter this trend, and that has received a lot of scholarly attention, is the use of so-called fact-checks; instruments which scrutinize (mis)information and indicate to what extent a specific piece of information is correct or incorrect. While many previous studies have demonstrated that fact-checks are effective tools in countering misinformation, less is known about the conditions under which fact-checks are the most effective, which is what we aim to examine with this study. Concretely, we look at two different factors. First, we are interested to what extent the source of the fact-check matters, comparing fact-checks coming from the traditional media - operationalized in our study as the public broadcaster – with fact-checks from independent organizations. Our expectation is that, in general, fact-checks coming from the public broadcaster will be more effective, as citizens are more familiar with this news source and hence may place more trust in it. However, we expect that this will be moderated by trust in the public broadcaster, both at the individual and country level. Second, we expect that the extent to which fact-checks are effective depends on how politicized they are, particularly when testing false claims made by political actors. We consider fact-checks to be politicized when they explicitly mention the political actor that made the (false) claim. Fact-checks are depoliticized when they do not mention which actor made the claim. Our expectation is that politicized fact-checks are less effective, especially among supporters of the political actor making the claim, as they may prime partisan identity and hence strengthen partisan motivated reasoning. Although we hypothesize politicized fact-checks to be less effective in all countries, we expect that this will especially be the case in more (affectively) polarized countries. To test our hypothesis we use data from an experiment that will be conducted across 16 different European countries, with mixed levels of media trust and polarization. In all countries respondents will read the same misinformation story with a national (right-wing) political actor making a false claim about the European immigration numbers. Subsequently, respondents will read a fact-check debunking this misinformation. The fact-check will differ on two dimensions, namely 1) whether it comes from the public broadcast or from an independent fact-check organization, and 2) whether it is politicized or depoliticized. Concretely, this means that we have a 2*2 factorial experimental design, along with a control group that only reads the misinformation article. Afterwards, respondent are asked to what extent they accept the misinformation claim, find the misinformation article credible, and to what extent they would share this article on social media. Hereby we take into account respondents’ prior opinion on immigration and support for the political actor that is used. Additionally, we study how country-level variables such as the level of media trust and polarization moderate the found effects.