To debunk or not to debunk health misinformation?

Submitted by gbianchi on Mon, 04/12/2021 - 15:42

The COVID 19 pandemic has seen an exponential production of information, some of it false or misleading, causing confusion and risk-taking behaviours. The information pandemic, or infodemic, can intensify or lengthen outbreaks . As facts, rumors, and fears mix and disperse, it becomes difficult to learn essential information about an issue.

Risk- and evidence-based analyses and approaches have been systematically employed by experts to manage the infodemic and restore trust in health authorities’ response.

A fundamental question among experts in infodemic management is how best to limit the impact of false narratives that may spread in a community. Should false claims be tackled head-on? If so, who is the best “messenger” for a debunk? Or should false claims be challenged indirectly through proactive counter-messaging that does not reference the false claim? In recent years, researchers trying to answer these questions have reached contrasting conclusions. In this post, we summarize a growing field of academic literature on the topic to distill key lessons from the academic debate and identify new opportunities for research moving forwards.

What is the backfire effect?

  • Researchers have inconsistently demonstrated two forms of “backfire effect” through experiments:
    • Familiarity Backfire:By repeating a false claim, debunks can enhance the familiarity of a false claim, leading to increased recall over time of the false claim.
    • Ideological Backfire:Direct factual contradictions cause backfire due to motivated reasoning processes.

What does the research tell us on the existence of backfire?

  • Backfire is not consistently observed, and initial findings of the effect have not been replicated in studies. 
  • There are factors that might exacerbate risk of backfire or reduce the efficacy of debunks:
      • Time between viewing of debunk and attempted recall
      • Repetition of fact-based messaging and repetition of false claims absent corrective information
      • Sequence of messaging (was debunk seen before the misinformation, i.e. was the misinformation novel?)

Are debunks effective?

What are the additional considerations for health communications regarding debunking and potential backfire?

  • Vaccine misinformation is a clear ongoing challenge to vaccine acceptance and vaccination campaigns (Dubé et al., 2015; Loomba et al., 2020)
  • Some research has suggested that around vaccines, for example, exposure to factual information among populations with high levels of vaccine skepticism led to higher rejection of vaccines (Nyhan et al., 2014; Pluviano et al., 2017; Pluviano et al., 2019)This research, however, focuses on a very basic presentation of “debunks” (i.e. “myth vs fact”), and itself highlights a challenge:                                
      • Therefore, we face a conundrum: while some accounts indicate that the best strategy to counter vaccine misinformation is to emphasize the facts instead of drawing further attention to false information to avoid a familiarity backfire effect, other accounts suggest that if a myth is not repeated when corrected, the associated lack of salience, conflict detection, and myth/correction co-activation may be equally or even more detrimental to belief updating than the boost of the myth’s familiarity (Pluviano et al., 2019)
    • In this context, where vaccine misinformation hinders acceptance, corrective information maycause backfire, and purely fact based information maylack salience - what information provision strategy can be pursued for optimal impact?

What are the risks in designing infodemic response strategy around backfire?

  • If the effect is overstated, then we are ceding ground and allowing harmful narratives to propagate unchallenged.
  • An overstatement of backfire has already provided cover for lackluster platform policy action. (Porter & Wood, 2020)

What research gaps can we address?

  • Research on backfire effect to date has seemingly been limited to specific debunking approaches that do not test nuanced debunking methodologies, which may encompass multi-platform efforts from diverse sources over the long term.
  • Research on backfire has been limited to mainly English-speaking contexts, and has not been carried out at scale with test groups in Africa, for example.
  • The impact of an existing relationship with an authoritative source (such as WHO, UNICEF, Africa CDC) on debunk effectiveness has yet to be researched.
  • Only limited publicly-available research has been carried out measuring the efficacy of fact checks in context(i.e. on a social media feed) - this research failed to find any backfire effect, and instead found increased awareness of accurate information. (Porter & Wood, 2020)

“We suggest that backfire effects are not a robust empirical phenomenon, and more reliable measures, powerful designs, and stronger links between experimental design and theory, could greatly help move the field ahead.” (Swire-Thompson et al., 2020)

Conclusion

There’s a clear need to better understand if and how debunks can be optimized for impact in different contexts, and to establish any observable backfire effects and how they can be mitigated. The research context is challenging, particularly given the multi-platform, multi-source, real-time nature of the misinformation challenge and the diverse strategies that can be adopted in an intervention. The Africa Infodemic Response Alliance, with its combination of fact checking expertise and behavioural science research capacity, has a unique opportunity to shed light on this crucial question for infodemic response.

 

Authors: 

AbdelHalim Abdallah, WHO Crisis Communications Officer for the African Region 

Tom Trewinnard, Fathm 

Bibliography

 

Summary - One Sentence
A fundamental question among experts in infodemic management is how best to limit the impact of false narratives that may spread in a community.
Language
English