Combatting Mis/Disinformation: Combining Predictive Modeling and Machine Learning with Persuasion Science to Understand COVID-19 Vaccine Online Discourse

Abstract

Health mis/disinformation can negatively impact health decisions and ultimately, health outcomes. Mis/disinformation related to COVID-19 vaccines has influenced vaccine hesitancy during a very critical time during the pandemic when globally, the vaccine was needed to attenuate the spread of the COVID-19 virus. This paper examines persuasive strategies used in Twitter posts, particular those with antivaccine sentiment. The authors developed a predictive model using variables based on the Elaboration Likelihood Model, Social Judgement Theory and the Extended Parallel Process Model to determine which persuasive tactics resulted in antivaccine, provaccine and neutral sentiment. The study also used machine learning to validate the persuasion variable algorithm to detect persuasion tactics in COVID-19 vaccine online discourse on Twitter. Understanding persuasive tactics used in antivaccine messaging can inform the development of a data-driven counter-response strategy.

Misinformation and COVID-19 vaccine uptake hesitancy among frontline workers in Tanzania: Do demographic variables matter?

“…Social judgment theory by Carolyn Sherif, Muzafer Sherif, and Carl Hovlan in 1960s 51 , 52 focuses on people’s values or issues within their latitude of acceptance. 38 , 53 The theory posits that persuasive messages are usually judged against an individuals’ existing position (anchor attitude). 47 If the new message falls into the latitude of acceptance, then it will be perceived as closer than it is to the anchor attitude.…”

“…Misinformation is not a new phenomenon in health. 38 Scannell et al 38 observe that health innovations and discoveries have been misrepresented and surrounded by myth and conspiracy throughout history. It was the same during the measles era, linking mumps, and rubella vaccine to autism.…”

“…It was the same during the measles era, linking mumps, and rubella vaccine to autism. 38 , 39 Studies 1 , 6–11 , 40 , 41 indicate that higher exposure to misinformation led to higher levels of negative attitudes toward the COVID-19 vaccine. Skepticisms such as vaccine will contain microchip; vaccine will cause infertility; vaccine will change deoxyribonucleic acid or DNA; vaccine will control the mind 42 or vaccine will cause blood clotting and eventual death 43 and so on remain.…”

A Comparative Case Study Analysis: Applying the HIPE Framework to Combat Harmful Health Information and Drive COVID-19 Vaccine Adoption in Underserved Communities

“…Customized search parameters within these tools help to identify the narrative trends, amplification within these trends, as well as if these narratives are being artificially amplified. A key element within the detect phase is the identification of specific persuasion tactics or triggers being used to drive amplification which utilizes HIPE™’s persuasion algorithm [ 25 ]. The persuasion tactics or triggers are grounded in health communication theoretical frameworks—Extended Parallel Process Model, Social Judgment Theory, and Elaboration Likelihood Model—which include elements of information processing, social factors, as well as self-efficacy [ 26 , 27 , 28 ].…”