Providing Normative Information Increases Intentions to Accept a COVID-19 Vaccine

"Unlike other ongoing, frequently observable preventative behaviors, like mask wearing, people may have little information about whether others intend to or have accepted a vaccine - which suggests messages with this information could have substantial effects."

There is evidence that people's preventative health behaviours are influenced by many social and cultural factors. Public health messaging can leverage the significant roles of social networks in shaping individual vaccination decisions. In fact, framing vaccination as a social norm has been suggested as an effective approach to building COVID-19 vaccine confidence. Using a randomised experiment (n = 484,239) embedded in an international survey, this paper investigates whether accurate information about descriptive norms - that is, beliefs about what other people do, believe, or say - can increase intentions to accept a vaccine for COVID-19.

Through a collaboration with Facebook and Johns Hopkins University, and with input from experts at the World Health Organization (WHO) and the Global Outbreak Alert and Response Network (GOARN), the researchers fielded a survey in 67 countries in their local languages, yielding over two million responses. This survey assessed people's knowledge about COVID-19, beliefs about and use of preventative behaviours, beliefs about others' behaviours and beliefs, and economic experiences and expectations. One finding was underestimation of vaccine acceptance by others, which could be partially caused by processes - such as news coverage of the challenges posed by vaccine hesitancy or diffusion of anti-vaccine messages on social media - that make hesitancy more salient. Beliefs about descriptive norms were in turn found to be positively correlated with vaccine acceptance. (See Related Summaries, below.)

Beginning in October 2020, for the 23 countries with ongoing data collection in the survey, the researchers presented respondents with accurate information based on how previous respondents in their country had answered a survey question about vaccine acceptance, mask wearing, or physical distancing. They randomised at what point in the survey this information was presented, which behaviour the information was about, and how they summarised previous respondents' answers.

On average, presenting people with normative information on the share of respondents in a country who will accept a vaccine increases stated intentions to take a vaccine. The effects on vaccination intentions are largely consistent across the 23 included countries but are concentrated among people who answered "Don't know" to the baseline vaccine acceptance question, consistent with the idea of trying to reach vaccine "fence-sitters". As a comparison point, these effects are over a third of the size of the total increase in vaccine acceptance from November 2020 to January 2021 across all 23 countries. Consistent with the hypothesis that this treatment works through revising beliefs about descriptive norms upwards, the researchers found significant effects of the normative information treatment in the groups that may be underestimating vaccine acceptance.

For mask wearing and physical distancing, the effects are smaller and often not statistically distinguishable from zero.

In addition to the primary experiment embedded in the global survey, the researchers conducted a supplementary survey in the United States (n = 1,350) over two waves (April 2 2021 and May 18 2021) to measure the link between vaccination intentions and self-reported vaccination uptake. This survey found that self-reported vaccination intentions are predictive of future, self-reported vaccination status. If respondents in the international experiment were to be vaccinated at the same rate as those in this supplementary analysis, there would be a 23.1 percentage point increase in vaccination rates among those who were unsure but were induced to say they would probably accept a vaccine and a 17.2 percentage point increase in vaccination rates among those who would probably accept a vaccine but were induced to say they would definitely accept a vaccine.

While in the randomised experiment norms are made salient through direct information treatments, the results have implications for communication to the public through health messaging campaigns and the news media. For example:

• If very high levels of vaccine uptake are needed to reach (even local) herd immunity and to minimise severe illness, it is reasonable for news media to cover the challenges presented by vaccine hesitancy, but the results suggest that it is valuable to contextualise such reporting by repeatedly noting the widespread norm of accepting COVID-19 vaccines.
• In an effort to influence the public, some public figures have documented receiving a COVID-19 vaccine in videos on television and social media. The positive effects of numeric summaries of everyday people's intentions documented here suggest that simple factual information about descriptive norms can similarly leverage social influence to increased vaccine acceptance.
• In addition to being salient, effective social norm interventions must be credible and not inconsistent with strongly held beliefs.

In short, results across countries suggest that accurate normative information often increases intentions to accept COVID-19 vaccines, and effects are largest in countries with higher norms. In addition, there is little evidence that providing normative information to those that overestimate vaccine acceptance results in decreased vaccination intentions. Taken together, this evidence suggests the positive effects from pro-social motivations and social conformity outweigh the possible negative effects from any free-riding on herd immunity.