2023/05/11 by Diego A. Reinero, Elizabeth Ann Harris, Steve Rathje +2 · 1 voice · 2 citations
Social Sciences · Computer Science · #Misinformation and Its Impacts #Media Influence and Politics #Hate Speech and Cyberbullying Detection
paper · pdf · doi:10.31234/osf.io/z4df3
Misinformation has become a global issue with potentially dire consequences (e.g., electionresults, COVID-19 vaccines, international conflicts). The most common strategy to reducebeliefs in misinformation is to issue corrections or fact-checks. However, fact-checks are oftenineffective for polarized issues and there has been debate over whether misinformationcorrections sometimes "backfire," causing people to become more entrenched in misinformation.Across three experiments (N = 1,217) we identify a key feature – the partisan identity of thesource – that explains when corrections will be effective and when they will be most likely tobackfire. We found that while corrections work on average, backfires still occurred 23% of thetime and were 52% more likely to occur when the correction came from a political outgroup vs.ingroup member – leaving people with more entrenched beliefs in misinformation. Indeed, theeffect of partisan bias on initial belief in misinformation was 5x larger than the effect ofcorrections. Partisan bias was most prominent and corrections were least effective when famouspoliticians were involved. In sum, corrections have small effects compared to partisan identitycongruence, and sometimes backfire–especially if they come from a political outgroup member.