Information Cascades and Bayesian Models in Election Misinformation
In the New York Times article, “Misinformation experts have been prepping for this day”, Tiffany Hsu details how several organizations have been taking measures to prevent the spread of misinformation during the midterm elections. We can connect election misinformation to two networks concepts: 1) information cascades and 2) Bayesian models.
In the article, experts from the Election Integrity Partnership described how the election would likely resurface misleading narratives and rumors from 2020. This dissemination of misinformation online can be described as an information cascade. As we learned in lecture, the requirements for an information cascade are that: people make decisions sequentially, everyone has private information, and people see what others do (but not what they know). In the case of social media election fraud accusations, each individual has private information about whether or not they witnessed election fraud. They can also post public information on social media claiming they witnessed election fraud, which may or may not be true. As people sequentially see posts on social media, they may choose to believe other accounts of election fraud over their own personal information. With how easily social media posts can go viral, this leads to the rapid spread of misinformation.
The article also describes how a notable difference with this year’s midterm elections is that a larger percentage of the public now expect misinformation online. In addition, many experts have warned the public to be skeptical of individual images or stories shared about polling incidents, as these reports are not verified by authoritative services, and can easily be taken out of context. This connects to our class discussion of Bayesian models. Bayes Theorem describes the probability of an event, given prior knowledge that might be related to the event. When we consider Bayes Theorem in the context of election misinformation, many individuals will have prior knowledge of potential misinformation during the 2022 midterms. This decreases the likelihood that they would believe election fraud posts, potentially breaking the information cascade. Overall, applying networks theories to election events sheds light on why misinformation can be so easily adopted on the Internet, and how we might be able to prevent this dynamic.
Resource: New York Times article, “Misinformation experts have been prepping for this day”, https://www.nytimes.com/2022/11/08/business/media/misinformation-midterm-elections