Mathematical Models Reveal How Election Misinformation Spreads Like a Virus

As the 2024 U.S. presidential election approaches, concerns about misinformation have reached alarming levels. Approximately 73% of Americans report encountering misleading election news, with about half struggling to distinguish fact from fiction, according to recent Pew Research data.

The comparison between misinformation and viral contagion is more than just metaphorical. Scientists have discovered that mathematical models originally developed to track disease outbreaks can effectively map how false information propagates through social networks. These findings suggest misinformation literally “goes viral” in ways that mirror biological pathogens.

Public concern about this phenomenon is widespread. A recent UN survey indicates 85% of people globally worry about misinformation—and for good reason. Foreign disinformation campaigns have grown significantly more sophisticated since the 2016 U.S. election.

The current election cycle has seen particularly dangerous examples: conspiracy theories about “weather manipulation” have undermined hurricane response efforts, false claims about immigrants eating pets have incited violence against Haitian communities, and election misinformation has been amplified by high-profile figures like Elon Musk, reaching potentially hundreds of millions of users.

Researchers are increasingly applying epidemiological models—specifically the “susceptible-infectious-recovered” (SIR) framework—to understand how misinformation spreads. These models, built on differential equations that measure rates of change, treat individuals as being in one of three states: susceptible to false information, infected with it, or recovered/immune to it.

“These mathematical approaches allow us to calculate crucial metrics like the basic reproduction number, or R0—the average number of people an ‘infected’ individual will spread misinformation to,” explains one researcher familiar with the work. Studies estimate most social media platforms have an R0 greater than 1, indicating fertile ground for epidemic-like spread.

A key advantage of these models is their ability to simulate potential interventions. For instance, studies show influential figures with large followings can function as “superspreaders,” similar to individuals who disproportionately spread disease during outbreaks. Election officials report being overwhelmed in their attempts to fact-check misinformation from these sources.

Simulations demonstrate that conventional debunking has limited effectiveness. In one model where exposure carries just a 10% chance of “infection,” populations exposed to election misinformation still show rapid growth in false belief adoption despite fact-checking efforts.

More promising is an approach called “psychological inoculation” or “prebunking.” This strategy involves preemptively introducing people to weakened forms of misinformation along with refutations—much like vaccines prime immune systems against future viral exposure. Recent research has used AI chatbots to generate prebunking content against common election fraud myths, warning people about tactics like claims of “massive overnight vote dumps flipping the election.”

Models show that when prebunking is deployed at scale, it can significantly limit the number of people who become misinformed, creating population-level resistance to false narratives.

“The point isn’t to suggest people are simply gullible disease vectors,” notes one expert. “But evidence clearly shows some fake news stories spread like simple contagions, infecting users immediately, while others act more like complex contagions requiring repeated exposure before beliefs change.”

These models can be adjusted to account for varying individual susceptibility across different populations. The approach is particularly valuable because most misinformation is diffused by relatively small numbers of influential superspreaders, mirroring viral transmission patterns.

While mathematical models have limitations, they offer powerful tools to predict misinformation spread and evaluate intervention strategies. Research examining social media dynamics during the 2020 presidential election validated this approach, finding that combined intervention strategies can effectively reduce misinformation’s reach.

As one researcher concluded, “Models are never perfect, but if we want to stop misinformation spread, we first need to understand it in order to counter its societal harms.”