Dynamic Effects of Falsehoods and Corrections on Social Media: A Theoretical Modeling and Empirical Evidence

Kelvin K. King, Bin Wang, Diego Escobari, Tamer Oraby

Research output: Contribution to journalArticlepeer-review

Abstract

Government agencies and fact-checking websites have been combating the spread of falsehoods on social media by issuing correction messages. There has been, however, no research on the effectiveness of correction messages on falsehoods and their dynamic interaction. We develop a theoretical model of the competition between falsehoods and correction messages on Twitter and show different interventions under which falsehoods could be hampered. Moreover, we use panel vector autoregressive models and machine learning techniques to empirically investigate the dynamic interactions between falsehoods and correction messages through a unique longitudinal dataset of 279,597 tweets. We find that correction messages cause an increase in the propagation of falsehoods on social media if their use is not optimized. This study highlights the importance of having government agencies, fact-checking websites, and social media platforms work together to optimize effective correction messages. We argue such an effort will counter the spread of falsehoods.

Original languageEnglish (US)
Pages (from-to)989-1010
Number of pages22
JournalJournal of Management Information Systems
Volume38
Issue number4
DOIs
StatePublished - 2021
Externally publishedYes

Keywords

  • combating fake news
  • fact-checking online
  • fake news
  • online corrections
  • online falsehoods
  • online misinformation
  • online rumor
  • panel vector autoregression
  • social media

ASJC Scopus subject areas

  • Management Information Systems
  • Computer Science Applications
  • Management Science and Operations Research
  • Information Systems and Management

Fingerprint

Dive into the research topics of 'Dynamic Effects of Falsehoods and Corrections on Social Media: A Theoretical Modeling and Empirical Evidence'. Together they form a unique fingerprint.

Cite this