When Corrections Fail: Effects of Misinformation Targets, Repeated Exposure, and Partisanship on Misinformation Beliefs

Yunya Song, Yuanhang Lu*, Stephanie Jean Tsang, Jingwen Zhang, Kelly Y. L. Ku

*Corresponding author for this work

Research output: Contribution to journalJournal articlepeer-review

Abstract

This study evaluates the effectiveness of 3 misinformation-correction approaches—fact-based, narrative-based, and literacy-based—in countering politically polarized misinformation. Using a 2 (misinformation target: protesters vs. police) × 4 (correction approach: no correction, fact-based, narrative-based, literacy-based) between-subjects online survey experiment with a representative sample from Hong Kong, we also examined how repeated misinformation exposure and partisanship influence responses to misinformation and corrections. Findings reveal that (1) none of the correction approaches significantly reduced misinformation beliefs, with no differential effects among them; (2) repeated exposure to reinforced misinformation beliefs, contributing to their persistence; (3) participants’ political affiliations shaped their beliefs in misinformation and corrections; and (4) exposure to partisan-incongruent misinformation increased acceptance of such misinformation. These results highlight the importance of considering political contexts and target sensitivity in misinformation correction strategies and underscore the need for tailored approaches, such as prebunking and media literacy, to build resilience against persistent misinformation.
Original languageEnglish
Pages (from-to)392-416
Number of pages25
JournalInternational Journal of Communication
Volume19
Publication statusPublished - Jan 2025

User-Defined Keywords

  • misinformation
  • correction
  • misinformation target
  • repeated exposure
  • partisanship

Fingerprint

Dive into the research topics of 'When Corrections Fail: Effects of Misinformation Targets, Repeated Exposure, and Partisanship on Misinformation Beliefs'. Together they form a unique fingerprint.

Cite this