As COVID-19 swept the country last spring, state and local officials responded swiftly. By late April, at least 319 million people in 42 states were subject to shelter-in-place orders aimed at containing infections, according to The New York Times.

The extraordinary measures raised a key question: As people became aware of their own risks, to what extent would they have limited grocery shopping, stopped dining out, or cancelled travel plans on their own? In other words, would it be correct to assume people would not have protected themselves but for drastic public health interventions?

A new Stanford study suggests that people’s perceptions of their own risks play an important role in their actions — and that shelter-in-place policies influence what they do, but not to the extent that some might think. The research, based on a May 2020 survey, finds that many people would have dramatically cut back on visiting movie theaters or getting a manicure, among other activities, without an order to stay home.

Researchers say the finding is significant for understanding how quickly people will return to pre-pandemic life and more broadly, how people’s perceptions of risk influence their behavior, and how much government policies have an affect on all of that.

“The bottom line is that people’s individual beliefs matter a lot and they ultimately drive behaviors,” says Maria Polyakova, an assistant professor of medicine at Stanford Health Policy, faculty fellow at the Stanford Institute for Economic Policy Research (SIEPR) and co-lead author of the paper. “You can’t assume that protective policies are a switch that you can turn on and off, and the economy goes up or down in response.”

Polyakova, an economist and assistant professor of health policy at Stanford Medicine, warns that, without taking into account how people behave in response to their own risk assessments, policymakers are likely to overestimate the effectiveness of shelter-in-place orders.

For instance, a failure to account for individual perceptions may explain why the nighttime curfews that some local governments had set last fall as coronavirus case numbers were rising were mostly ignored, says Polyakova.

The research, released Monday by the National Bureau of Economics Research as a working paper, was also led by M. Kate Bundorf, who was a SIEPR senior fellow when the study began and is now a professor of public policy at Duke. Other collaborators were Grant Miller, a Stanford professor of medicine and SIEPR senior fellow; Jialu Streeter, a research scientist at the Stanford Center on Longevity; and Jill DeMatteis and Jonathan Wivagg, both of Westat, which facilitated the survey.

Perceiving Risks — Hit and Missed

Disentangling shelter-in-place orders from people’s actions based on their own sense of the risks isn’t easy, says Miller. Tracking movement based on smartphone GPS wouldn’t say much about how government mandates altered perceptions of risk, and attempting to draw meaningful comparisons between two or more states is difficult because attitudes from one jurisdiction to the next can differ significantly for a host of reasons, including political ones.

Polyakova and her team came up with another idea: To conduct a national survey that asked people to disclose how they changed their behaviors in response to a shelter-in-place order and what they would have done in the absence of one. It’s not an ideal experiment, in that people may not accurately report what they would do in a hypothetical scenario. But even with that limitation, a survey could generate useful insights.

In the end, the researchers analyzed the online responses of 1,127 individuals, all of whom were subject to a stay-at-home order at the time (and received $6 for completing the survey). The pool of participants was diverse, with 49 percent reporting as male, 16 percent Hispanic, 9 percent non-Hispanic Black, 31 percent with at least a bachelor’s degree, 28 percent as essential worker, and 55 percent as having a chronic health condition. Respondents varied, too, by socioeconomic status and geography. (The complete dataset is here.)

To start, the researchers needed to understand how accurately people understood COVID-19’s risks — and how they changed their behavior under a stay-at-home order and how they thought they would have responded without one.

On the first point, the researchers found that people hugely overestimated the absolute risk of contracting COVID-19 and of becoming seriously ill as a result. While data showed that, as of the end of May 2020, the prevalence of COVID-19 in the United States was just 0.5 percent, the respondents estimated they had between a 40 percent and a 62 percent chance of catching the virus, respectively, at the grocery store or while sharing a ride. The results fell in line with other extensive research showing people often miscalculate the threat of small and highly publicized risks.

Survey participants on average were much better, however, at assessing their own relative risks. Hispanics, Blacks, individuals with pre-existing health conditions, and those living in COVID-19 hot spots correctly recognized they were at higher risk of contracting the virus. They also, along with the elderly, knew that they were more likely to get seriously ill.

Aligning Policies, Perceptions

Next, the researchers looked at the extent to which people reported cutting back on activities under a stay-at-home mandate and how much they thought they would without one.

Of the respondents, about 40 percent said they had reduced their grocery shopping by “a lot,” and 79 percent had similarly reduced their dining out. Even without a government directive, they said on average that they were likely to limit these and other activities — not as much, but still by a significant amount. The higher the risk they perceived an activity like eating at a restaurant, getting a haircut or going to the gym, the more likely they were to avoid it.

One surprising exception were individuals who considered themselves at high-risk for serious illness or even death if they contracted COVID-19. Those individuals reported that they were more likely to go to the grocery store or even the movies once stay-at-home orders were imposed. While the study did not examine the reasons for that response, Polyakova thinks it’s because the restrictions helped guide this group as to how much risk they really faced. By severely limiting what the overall population might do (for example, attending parties), the policies could have given high-risk people the (correct) perception that venturing outside under certain circumstances was consequently safer.

“If you believe you’re at a greater risk of a serious complication and there is no government intervention, then you may not leave your home at all,” Polyakova says. But shelter-in-place policies had the effect of equalizing people’s behaviors, she says, making certain high-risk individuals more willing to leave their homes and low-risk people less likely to do so.

Overall, the researchers say it’s important for public health officials to take into account people’s perceptions of their personal risks when issuing mandates — and evaluating their effectiveness. And now, as economies re-open, policymakers should consider designing information campaigns to better educate constituents about actual risks and how policies are intended to address them.

“Our research suggests that risk communication is an important policy tool and that people will take into account the specifics of their personal situation in responding to that information,” says Bundorf, the co-lead author. “Our results also point to the importance of considering how policies affect people with different beliefs when studying the impact of policies.”