Learning about the opposing party’s vaccination rate influences health choices

Public health campaigns often rely on broadcasting facts and statistics to encourage healthy habits. People are more likely to express an intent to receive a COVID-19 vaccine after learning that vaccination rates among members of the opposing political party are higher than they previously thought. The findings, published in PLOS One, suggest that public health campaigns highlighting low vaccination rates to urge action might inadvertently discourage people from getting immunized.

In the United States, political affiliation heavily influences healthcare choices. Currently, whether a person identifies as a Democrat or a Republican is a stronger predictor of their COVID-19 vaccination status than their education level, race, or health insurance coverage. Past research indicates that people often judge the severity of the pandemic and the necessity of government interventions through a partisan lens.

Public health officials frequently attempt to change behavior by providing the public with new data. The success of this strategy heavily depends on what the audience already believes. If a person thinks a virus is exceptionally rare, hearing that it actually infects one in five people might prompt them to take precautions. If that same person already assumed the virus infects half the population, the identical statistic might make them feel safer and less likely to act. Researchers refer to this gap between an individual’s preconceived belief and the actual statistic as an information shock.

Researchers Brandyn F. Churchill of American University, along with Xiaoxue Sherry Gao and Rong Rong of the University of Massachusetts Amherst, wanted to test how these information shocks interact with political identity. They specifically wanted to see how learning new facts about the health behaviors of opposing political groups shapes personal medical decisions.

To establish a baseline of actual COVID-19 experiences, the researchers first conducted an initial survey of 200 adults on an online research platform in May 2024. The group consisted of 100 self-identified Democrats and 100 self-identified Republicans. The participants answered questions about their recent experiences with the virus, including whether they had contracted it or sought medical care. They also reported whether they had received a vaccine, experienced side effects, or vaccinated primarily to protect others.

Free daily newsletter

The initial survey revealed that both political groups experienced similar rates of infection. Both groups also sought medical care for the virus at similar frequencies. There was, however, a vast divide in vaccination behavior. Nearly 46 percent of the Democrats reported receiving a recent COVID-19 vaccine, compared to only 28 percent of the Republicans.

Building on these baseline figures, the researchers designed a main experiment in August 2024 involving 579 different participants. This group included 288 Democrats and 291 Republicans, representing a mix of ages and backgrounds.

The researchers asked these individuals to estimate the responses given by the people in the initial survey. Half of the participants were asked to guess the survey answers of their own political party. The other half were asked to guess the answers of the opposing political party.

To ensure participants answered honestly rather than simply expressing a political bias, they were offered a financial bonus. The participants placed digital bets on how a randomly selected person from the first survey had answered the health questions. If their estimates closely matched the actual data, their chances of winning a monetary reward increased.

The estimates revealed widespread misunderstandings about the realities of the disease. Across the political spectrum, participants overestimated how frequently people caught COVID-19 by about 20 percentage points. They also overestimated how often the virus required medical intervention by 22 percentage points.

Both Democrats and Republicans tended to underestimate the prevalence of vaccine side effects. They also underestimated how many people chose to get vaccinated specifically to protect others.

Participants were no better at accurately estimating the behaviors and experiences of their own political party than they were at estimating those of the opposing party. Both groups consistently overestimated how many Republicans were vaccinated against the virus.

After the participants locked in their estimates, a computer program randomly selected one of the five health topics. The program then revealed the actual statistic from the initial survey to the participant. By comparing the participant’s original guess to the newly revealed statistic, the researchers could measure the size and direction of the information shock. Finally, the researchers asked the participants if they intended to receive an updated COVID-19 vaccine within the next twelve months.

Learning the true statistics regarding infection rates, disease severity, or side effects did not alter the participants’ stated plans to get a vaccine. Seeing the true vaccination rates of their own political party also produced no changes in their intentions.

When participants learned the actual vaccination rate of the opposing political party, their behavior shifted. Participants who realized they had underestimated the opposing party’s vaccination rate became more likely to state an intention to get vaccinated themselves. Those who learned they had overestimated the opposing party’s vaccination rate became less likely to seek a vaccine.

The researchers conducted mathematical tests to ensure this relationship was not a statistical fluke. When they swapped the data to pretend the participants had seen a different health statistic, the effect disappeared. When they looked at whether the information changed past vaccination behavior, there was no connection. This adds weight to the idea that the specific information about the opposing party’s current vaccination rate drove the change in their future intentions.

The researchers attribute this pattern to a social pressure model. In this framework, people are more inclined to adopt a health measure when they see others adopting it. When people learn that a group they view as generally opposed to a behavior is actually participating in it, they may feel increased social pressure to follow suit.

The results present a potential hazard for modern public health messaging. Health agencies often try to motivate the public by emphasizing that vaccination rates are falling below optimal levels. Because people generally overestimate how many of their peers are vaccinated, broadcasting low compliance rates serves as a negative information shock. Hearing that few people are getting the shot could reduce social pressure and inadvertently depress vaccination numbers even further.

While the experiment highlights shifts in stated intentions, an expressed intent to receive a vaccine does not guarantee that a person will actually schedule an appointment and get the shot. The researchers did attempt to follow up with the participants four months later to ask if they had received a vaccine. Many of the participants did not respond to the follow-up survey, meaning the researchers could not reliably track how often intentions translated into real-world behavior.

The study design also limits the ability to explain the underlying psychological mechanisms. The data show that participants reacted to the habits of the opposing political party, but the study does not reveal exactly why this specific information proved more persuasive than data about their own peers. The overall number of participants also constrained the researchers’ ability to analyze differences across more specific demographic categories, such as age, race, or income brackets. Future studies will need to investigate the cognitive pathways that make discordant political information so influential in personal medical decisions.

The study, “Partisan differences in healthcare decision-making: Evidence from a vaccine experiment,” was authored by Brandyn F. Churchill, Xiaoxue Sherry Gao, and Rong Rong.