To prevent the spread of COVID-19, people have been encouraged to adopt preventative measures such as hand-washing and mask-wearing. Through a survey of over 200,000 Facebook users in 26 countries conducted on Facebook between May and October 2020, we examined how the adoption of such precautions has ebbed and flowed over time, and how precaution-taking relates to both offline activity (e.g., seeing people in person) and online activity (e.g., viewing COVID-19-related content on Facebook).
In this survey, we asked about whether people adopted any of seven precautions:
From May through October 2020, hand-washing and mask-wearing were the two most common precautions that people took worldwide. Of the people we surveyed, 73 percent reported washing their hands or using hand sanitizer, and 72 percent reported wearing a face mask.
Figure 1: Hand-washing and mask-wearing were the most commonly reported precautions that people took between May and October 2020 because of the COVID-19 pandemic. Country flags indicate prevalence in the respective countries (Taiwan, Peru, and the United States).
Figure 1 also shows how these overall rates compare with those of three countries: Taiwan (0.29 deaths per million as of early October 2020), the United States (640 deaths per million, and Peru (1,000 deaths per million). While people in Taiwan reported lower rates of following all precautions, people in Peru reported the highest rates of the more restrictive precautions (such as not leaving home at all).
Despite the COVID-19 case rate rising and the death rate holding relatively steady from May to October 2020, likely through a combination of improved testing and treatment, people took fewer precautions over the same time period (Figure 2). In May 2020, three in five people reported avoiding public locations; by October 2020, only two in five people did.
Figure 2: Globally, between May and October 2020, people were less likely to take precautions over time.
Precaution-taking decreased over time. For all precautions surveyed except for mask-wearing, the proportion of respondents who reported taking them decreased over the period of the study. Mask-wearing increased between May and September 2020 but decreased in October 2020. In the United States, there were significant decreases in people washing their hands, avoiding crowded places, avoiding public locations, and not leaving home except for essentials.
Women took more precautions than men (Figure 4). This aligns with recent research that women were more likely to perceive COVID-19 as a serious problem and to adopt preventive measures.
Figure 3: Women took more COVID-19-related precautions than men.
Age differences. Older adults took more precautions, primarily in hand-washing and social distancing (Figure 4), while 18- to 30-year-olds were most likely to not take any precautions at all. These trends are largely consistent with the increase in risk of getting COVID-19 as one gets older (e.g., the case fatality rate is about 10 times higher for those above 60 than it is for those under 40).
Figure 4: Older adults took more precautions than younger adults.
Precaution-taking differed substantially by country. Not only were there large differences across countries in measures taken, such as mask-wearing and social distancing, but depending on the country, related precautions (e.g., mask-wearing vs. social distancing) were adopted at vastly differing rates.
Differences in mask-wearing (Figure 5). A majority of Americans wear masks: On average, three out of four people reported wearing a mask as a precaution. People in Japan were most likely to wear masks (87 percent), while Swedes were least likely to wear masks (7 percent), likely because Sweden’s public health authorities discouraged people from wearing masks. And though only 68 percent of people in Taiwan reported wearing a mask, Taiwan has one of the lowest death rates in the world, likely because of the successful countermeasures taken from the beginning of the pandemic.
Figure 5: Almost nine in 10 people in Japan reported wearing a mask as a preventive measure, while fewer than one in 10 people in Sweden reported wearing a mask.
The relatively high compliance in the United States is notable, despite poorguidance from the federal government during the time. Mask-wearing was highest in Maryland (93 percent) and lowest in Tennessee (58 percent) (Figure 6). These largely correspond to whether states had instituted a mask-wearing mandate. For example, Tennessee’s governor did not support a mask mandate. These findings on mask-wearing largely mirror existing surveys (e.g., by YouGov). Still, even higher rates of mask-wearing may be necessary to substantially reduce the spread of COVID-19, as described in a recent CMU blog post.
Figure 6: Between May and October 2020, in the United States, mask-wearing was highest in Maryland (MD) and lowest in Tennessee (TN). Grayed-out states had fewer than 50 responses.
Social distancing. People in Canada, the United Kingdom, and the United States reported some of the highest rates of social distancing in the world (Figure 7). Notably, 70 percent of Swedes reported social distancing, despite only 7 percent wearing masks. Like the mask-wearing rate, the social distancing rate was lower in Taiwan (27 percent), likely reflecting the already-low prevalence of COVID-19.
Figure 7: People in Canada, the United Kingdom, and the United States reported some of the highest rates of social distancing. In contrast with Swedes’ low rate of mask-wearing, they reported a relatively high rate of social distancing.
Precaution-taking was correlated with the deadliness of the disease. People reported being more cautious when the risk of dying from COVID-19 was higher. In countries with more COVID-19 deaths per million, people were more likely to take one or more precautions, and this trend was strongest for social distancing in particular (Figure 8).
Figure 8: Social distancing was correlated with the death rate in each country.
Surprising differences. While one may expect the adoption of different precautions to be largely correlated with each other, there were some countries where some precautions were relatively more prevalent than others:
Figure 9: There are substantial differences in how common different preventive measures are, both within and across countries. Blue indicates relatively higher prevalence of a precaution in a country compared with other countries, while red indicates relatively lower prevalence.
Why do these country differences exist? Multiple factors play a role: the death rate, perceptions of risk, preexisting norms (e.g., Japan’s culture of mask-wearing), and differences in government directives (e.g., Sweden discouraged mask-wearing). Other factors such as population density (e.g., it may be harder to avoid crowds in cities) also likely play a role.
In the previous section, we described how precaution-taking varied by the type of precaution, a person’s age and gender, and the country that they live in. In this section, we explore how precaution-taking related to whether people saw others in person, and how it related to generally feeling better or feeling worse.
COVID-19 is a communicable disease, so many precautions are aimed at reducing the risk of person-to-person transmission. As Figure 10 shows, people who adopted “protective” measures (e.g., mask-wearing and hand-washing) reported having interacted with more people offline outside of their homes, while those who adopted “isolating” measures (e.g., staying home) reported having interacted with fewer people offline.
Figure 10: People who adopted more “protective” measures (e.g., hand-washing) had more offline interactions; people who adopted more “isolating” measures (e.g., not leaving home) had fewer offline interactions. Error bars represent 95 percent confidence intervals; darker bars indicate statistically significant associations (p < 0.05).
Does taking more precautions lead to more offline interactions (or vice versa)? On one hand, some researchers have hypothesized that precautions such as mask-wearing may give people a false sense of security, leading them to become less careful in their interactions with others (e.g., how often a person interacts with others outside their household). On the other hand, people with more offline interactions likely have a stronger incentive to adopt “protective” precautions.
Our findings show little evidence of the former. When we examined the extent to which precaution-taking in the past was correlated with future offline interactions, we found that taking more precautions was associated with a small subsequent increase in offline interactions in Spain and a decrease in offline interactions in Indonesia and the Philippines. Additionally, in Canada and Japan, having more offline interactions was associated with subsequent decreases in the number of precautions people took. More research is necessary to understand and verify these observations.
Some studies have warned that “quarantine and social distancing […] lead to elevated levels of loneliness and social isolation.” Past research found that loneliness is associated with negative affect, so we expected largely similar findings with respect to precaution-taking. However, this was not the case.
People who took precautions reported feeling better. To understand whether people felt better or worse, the survey included a question about the extent to which a person “felt good most of the time” and a question about the extent to which they “felt bad most of the time.” People who reported taking precautions were more likely to feel good most of the time (Figure 11) and less likely to feel bad most of the time (not shown). Only in the extreme case of not leaving home at all did people report feeling worse.
Why might people who take precautions feel better? They may feel safer from taking these precautions or feel better from actively playing a role in their community to stem the spread of the disease. Taking “protective” precautions such as mask-wearing may also have allowed people to better maintain social connections with each other and thus feel more supported.
Figure 11: Precaution-taking is generally associated with feeling better.
In this section, we show how seeing COVID-19-related content on Facebook may be related to precaution-taking, and how precaution-taking may be related to use of the “care” reaction.
People who saw more COVID-19-related content on Facebook were more likely to take one or more precautions. After adjusting for differences in demographics and on-Facebook activity (e.g., the total number of posts that a person saw), we found a significant association between the proportion of COVID-19-related content seen and whether any precautionary measures were taken. On one hand, people who see more COVID-19-related content may end up becoming more informed and thus take more precautions; on the other hand, people who are more concerned about COVID-19 may also be the people who follow and view more COVID-19-related content.
Which precautions were associated with viewing more COVID-19-related content? People who viewed proportionally more COVID-19-related content were more likely to wear a mask (Figure 12). They were less likely to not leave their homes at all.
Figure 12. People who viewed proportionally more COVID-19-related content were more likely to also report wearing a face mask or washing their hands.
Mask-wearing was associated with using the “care” reaction more. One commonly emphasized benefit of mask-wearing is that it protectsothers from being infected by the wearer. This benefit is largely different from many of the other precautionary measures, where the perceived benefit accrues largely to oneself. As such, mask-wearing may be more strongly associated with caring for others compared with other precautionary measures. To understand if this may be the case, we examined how people used the “care” reaction, which typically indicates support or care for others.
In regression analyses, mask-wearing was associated with proportionally greater use of the “care” reaction (Figure 13). This corroborates prior work that indicated that antisocial traits such as a lack of empathy and risk-taking were associated with a lower likelihood of mask-wearing or social distancing. While there were also significant associations with respect to not leaving home or avoiding crowded places, follow-up analyses indicate that these were because mask-wearing was correlated with both.
Figure 13. Mask-wearing was associated with using the “care” reaction more, suggesting a link between mask-wearing and empathy.
In summary, from May through October 2020, people took fewer preventive measures against COVID-19. Across 26 countries, including the United States, hand-washing and mask-wearing were the two most common preventive measures. Women adopted more precautions than men, and older adults adopted more precautions than younger adults.
Still, there were differences in how often precautions were taken in different countries. While Japan had the highest rate of mask-wearing (88 percent), social distancing was far less prevalent (39 percent). Conversely, Sweden had the lowest rate of mask-wearing (7 percent), but social-distancing was more common (71 percent).
And while some precautions were associated with seeing more people in person (e.g., wearing masks), others were not (e.g., avoiding public locations). Precaution-taking was also associated with feeling better.
Additionally, people who saw more COVID-19-related content on Facebook were also more likely to take one or more precautions. Mask-wearing was also associated with using the “care” reaction more, so people who wear masks may tend to be more supportive of others.
We hope that these findings begin to shed light on the relationship between COVID-19, the precautions that people take, and the online and offline interactions that people have. For more information about what Facebook is doing to keep people safe and informed about the coronavirus, read the latest updates on Newsroom.