Implicit in the discussion and consternation around vaccine hesitancy has been a wish: If only there were a message that could break through skepticism and persuade unvaccinated people to get their shots. What a world that would be.

But over the past year, this really did happen. While there are still too many holdouts, many people who were skeptical or hesitant changed their minds and got their Covid-19 shots.

A new study by Vincent Pons at the Harvard Business School and Vincenzo Galasso and Paola Profeta at Bocconi University in Milan may help explain why. They found that people can in fact be persuaded by the right message, offering new evidence about the best ways to reach skeptics. Separate analyses have identified other factors, such as seeing other people be safely vaccinated and hearing about the benefits of vaccines from doctors and loved ones.

This study assessed which messages were most effective in increasing people’s willingness to get vaccinated, lessons that could still help advance the goal of vaccinating more people against Covid-19. There is a range of vaccine skepticism — from people who wanted to wait and see before getting their own shots, to the people uncertain about booster shots, to the vaccinated parents of unvaccinated but eligible children. Successful messaging could help in winning them over and may be instructive for future public health campaigns as well.

The researchers surveyed more than 6,000 people in the United States, United Kingdom, European Union, Australia, and New Zealand. They contacted them first in December 2020, to assess their intentions before the vaccines were widely available, and asked them to pick a number between 0 and 10 to represent their likelihood of getting vaccinated. Then they followed up in summer of 2021 to see how people actually behaved.

To me, their most interesting findings concerned the most ardent vaccine refusers. Six months later, one-third of the people who had rated themselves 0 in December had gotten vaccinated.

So what happened? What convinced them?

Some of it was circumstances. Among those who had put themselves between 0 and 3 on getting vaccinated, those who were older (and therefore at higher risk of serious illness) and concerned with their health risks were more likely to get vaccinated in spite of their skepticism. So did the people who anticipated indirect exposure to Covid-19 through their friends or relatives. People who consumed more traditional media and who had more trust in scientists were also more likely to come around.

Vaccine mandates were not in effect at the time of these surveys, but the study generally found a mixed response to compulsory vaccinations among the respondents.

The researchers experimented with four messages in December to see how they would move people’s intentions then and their actual behavior six months later:

1. Self-Protection (If you get vaccinated, you could avoid getting infected)
2. Protecting Others (If you get vaccinated, you could avoid passing the virus to others)
3. Protecting Health (If you get vaccinated, it can help protect the health of people in your country)
4. Protecting the Economy (If you get vaccinated, it can help a return to economic activity and reduce unemployment)

A control group heard no messaging at all.

The self-protection message had a small positive effect in increasing people’s vaccine intentions, but it was the smallest effect of the four and was not considered statistically significant. The altruistic messages — Protecting Others or Health or the Economy — had a more meaningful impact on people’s intentions immediately afterward.

The same pattern bore out six months later in behavior, though there was some variation across countries. Protecting Health ended up most effective in the EU, while Protecting Others or the Economy proved more persuasive in the US and UK. The researchers argued that, because this line of questioning reflected people’s actual decisions, not only their intentions, it should be less susceptible to social desirability bias (i.e., people saying what they think they are supposed to say).

Those results were for the whole population, from the very pro-vax to the very anti-vax. But what worked best with the toughest people to win over?

Researchers found a positive and significant effect with the Protecting Health message in increasing the vaccination rate among anti-vax people in EU countries; there was also a positive but not statistically significant effect of the Protecting the Economy message on the vaccination rate of anti-vax people in the US and UK.

The latter point might sound discouraging, but the authors include a notable caveat. The people whom the study attempted to persuade and still refused to get vaccinated may be the most dedicated anti-vaxxers, even more resistant than anti-vaxxers in the control group who did not hear one of the tested messages.

The authors summarized their findings like this (emphasis mine):

Our most striking result is that informational treatments provided in the first wave affected not only vaccination intentions expressed in the same survey but also actual vaccination rates recorded six months later. Our messages even increased vaccination among individuals who had initially expressed anti-vax attitudes. Overall, altruistic messages had the largest effect. The important differences in the relative effectiveness of our different treatments across countries indicate that future information campaigns should be tailored to the context to be most impactful.

Translation: People could be persuaded to get vaccinated. It is not a lost cause. But we should be mindful that different messages work best with different people and construct our communications campaigns accordingly.

One wrinkle that bears further exploration: Among those who were least likely to change their minds were the people who are generally more risk-averse. The study’s authors speculated that “these individuals may be more concerned about possible negative side effects of vaccination than about the risk of getting Covid-19.”

That theory would track with something I heard in my interviews with pediatricians who have been talking with parents about getting their children vaccinated. Parents would have misperceptions about the relative risks of vaccination versus contracting Covid. To give one example, they might be aware that the vaccines carry a small risk of heart inflammation but might not be aware that Covid-19 itself brings a much higher risk of the same condition. Figuring out ways to better communicate the comparative risks of vaccination (and other interventions) versus infections is one takeaway from the struggles to communicate with the public during the pandemic.

It is tempting to think of vaccination intentions as static. You will or you won’t, for whatever reason: ideology, trust in institutions, personal assessment of your risk. And those variables have indeed been instrumental to people’s choices about getting the Covid-19 vaccines.

But people can be moved, too, according to these findings — not necessarily by emphasizing their personal health, but by persuading them that they will do good for the people around them and the country at large.