Reports

Earlier this month, we witnessed a vigorous back and forth between the White House and Facebook regarding the role of the company in circulating health misinformation. On the White House side, the Surgeon General released an Advisory declaring misinformation a public health threat to the nation, highlighting, in part, the role of social media. This was followed the next day by a pointed statement by President Biden-- “They’re killing us!” Facebook responded with a blog entry, Moving Past the Finger Pointing, asserting that it was not responsible for the plateauing vaccination rates in the US. The company pointed to the fact that 85% of Facebook users in a large survey by Carnegie Mellon (supported by Facebook) reported being vaccinated or wanting to be vaccinated. 

Major gender and age gaps remain, with young mothers and mothers of young children being the most resistant to vaccinating their kids and requiring COVID-19 vaccination for in-person school attendance. 

● Among older mothers, vaccine resistance is declining and support for school vaccine mandates is increasing. 

● Black parents have become substantially less resistant to vaccinating their kids. However, among Democrats, Black parents remain by far the most vaccination-resistant group, followed by Hispanic Democrats, then White Democrats, due in part to lower levels of trust in government and healthcare institutions. Asian American parents are the most pro-vaccine among both Democrats and Republicans. 

● Support for school vaccine mandates has grown substantially from 54.4% in the winter to 61.3% in the summer. Support has grown among both Republicans and Democrats, though a major partisan gap 

The COVID-19 Consortium for Understanding the Public’s Policy Preferences Across States 6 

persists. Overall, Americans who are more liberal, educated, higher income, and urban are more likely to support school vaccine mandates. ● Among parents specifically, every group except young mothers and mothers of young children has become more supportive of requiring vaccination to attend school. 

Overall, the levels of social isolation have increased since the first waves of the pandemic, in spring, 2020, but have been showing signs of improvement since January 2021. After its initial increase, the percentage of socially isolated respondents declined most among those with high income and education. We also observed a relatively faster drop in isolation among religious and older Americans. Conversely, unemployed and low SES respondents have barely recovered from the increases in social isolation suffered earlier in the pandemic. In addition, different types of support are not equally available across demographics. Men are substantially more isolated than women with respect to emotional support (but about equal in terms of economic); and white respondents are more socially isolated with respect to economic support; but less isolated emotionally relative to other racial/ethnic groups. Finally, we find a strong association between social isolation (specifically, in terms of relationships available for emotional support) and moderate to severe depressive symptoms. 

● We do not see evidence of an improvement in the prevalence of depression and other measures of mental health, despite marked improvement in the pandemic in the United States since the winter. 

Overall, 28% of those surveyed report levels of depression that would typically trigger a referral for evaluation and treatment; these numbers have diminished slightly compared to their peak of 30% in December 2020 but remain about three times those observed in the pre-COVID era, and elevated compared to the low of 25% in late June 2020. Rates of anxiety also remain elevated at 25%, but down from 28% in December 2020. Notably, 23% of respondents still endorse thoughts of suicide at least occasionally, similar to the rates observed in December. 

● Young adults continue to be the hardest hit, with 42% meeting at least moderate depression severity criteria, followed by 25-44 year olds at 32% and 45-64 year olds at 20%. Respondents age 65 and older are the least impacted, on average, with 10% meeting criteria for moderate or greater depression. 

● Parents with children at home have consistently experienced elevated rates of depression, with a gap of about 10% between parents and nonparents - currently 35% versus 25%. Some of this difference likely reflects age - that is, parents are generally younger than nonparents. The remainder may reflect such additional stresses as remote education. 

● Respondents without college education continue to report higher levels of depression (30%) than their college educated counterparts (23%). 

● Similarly, those earning less than $30,000 per year continue to demonstrate the highest rate of depression (35%) - but the pattern among those earning more is less clear, with lowest levels of depression reported among those earning $50-150,000 per year (23-24%). 

● The rate of depression among Asian American respondents has increased since March 2021 from 22% to 27%; while these numbers have generally been lower than any other racial/ethnic group, they are now similar to all except Latino respondents. 

“Vaccine certification” requirements were first introduced in the United States in response to smallpox in the 1800s. Today, physicians routinely provide certification that students have received an array of vaccinations that are required in order to attend school or summer camp, or to participate in sports or other group activities. A similar concept -- recently labeled “Vaccine Passports” (“VPs”) -- has been suggested with respect to COVID-19 vaccination, requiring individuals to provide proof of vaccination in order to engage in certain activities, such as indoor dining or shopping, flying on an airplane, or attending school. However, requiring that Americans prove that they have been vaccinated in order to participate in such activities has generated fierce opposition. The success or failure of these initiatives likely rests on a better understanding of the breadth and depth of this opposition, and its basis. Is it simply that the term “vaccine passport” has been politicized? Or is the underlying concept viewed by Americans as problematic? Or, in the wake of a year of lockdowns and mask mandates, are Americans concerned about yet another government mandate? 

The purpose of this report is thus to explore two questions. First, is public resistance to vaccine passports mainly due to the politicization of the term itself, or is the public genuinely opposed to any sort of vaccine mandate? Second, do those who oppose vaccine passports object only to the government requiring them, or do they object to any institution, public or private, making them a requirement? 

● A majority of Americans (62%) support “Requiring everyone to get a COVID-19 vaccine.” Even higher percentages support mandatory vaccines “to get on an airplane” (67%) and “to go back to university” (65%). 

● A majority of respondents in every state support some form of vaccine requirement. Air travel requirements are the most popular, with majority support in all 50 states, followed by university requirements (supported by majorities in 49 states), and finally requirements for children to return to school (majority support in 43 states). Men are more supportive of each potential vaccine requirement than women. This mirrors the gender divide we see in vaccine hesitancy (see Report #48) and willingness to vaccinate one’s own children (see Report #45). 

● Higher income, more highly educated, and non-White (particularly Asian American) respondents show greater support for vaccine requirements. Nevertheless, a majority of people in every racial, educational, and income group supports requiring “everyone” to get vaccinated, as well as supporting mandatory vaccines for boarding airplanes, children attending schools, and returning to universities.

●Despite widespread efforts by Republican politicians to mobilize opposition, more than four in tenRepublican respondents support requiring everyone to get vaccinated (42%). Support among Democrats is nearly twice that (81%).

While, on average, adherence to health and safety guidelines has slightly decreased since the spring, spikes in caseloads near the end of 2020 may have contributed to higher rates of social distancing, safer behaviors outside of the home, and stricter following of public health recommendations in California. Our survey data suggest that since the start of the new year, as more Californians have been getting vaccinated, they have begun to relax some of their adherence to social distancing and other public health recommendations. 

We evaluate whether the temporary pause of the Johnson & Johnson vaccine on April 13 negatively affected vaccine attitudes. We find no evidence that the pause affected vaccine intentions. However, the rapid rate of vaccinations has largely exhausted the number of vaccine enthusiastic people, likely driving the current slow down in vaccination rates. 

In our Report #48 from April 2021, we showed findings from our most recent survey indicating that the FDA and CDC’s recommendation to temporarily pause the Johnson & Johnson vaccine did not affect overall trends in intentions to vaccinate in the United States, despite a majority of respondents indicating that they were aware of the decision. Here, we briefly elaborate on those findings using a multivariate regression. 

● Since February, the gap in attitudes between fathers and mothers has widened. While fathers became marginally less resistant, falling from 14% to 11% since February, over a quarter of mothers still say they are “extremely unlikely” to vaccinate their children. 

● Educational, income, and partisan divides in childhood vaccination attitudes have become more pronounced. Parents in households making less than $25,000 per year, parents without a college degree, and Republican parents have become more resistant to vaccinating their children. Resistance has decreased for college-educated, high income, and Democratic parents. 

● Parents of teenagers are less resistant to having their children vaccinated than parents of small children, and—in the case of mothers—slightly more supportive of school vaccination requirements. 

● Support for school vaccination requirements has grown slightly from 54% to 58%. This increase holds for most gender, race, and income categories. However, among Republicans, support remains virtually unchanged. 

● Mothers are less likely to support school vaccination requirements than other women, while fathers are more likely to support school vaccination requirements than other men. 

● Among youth who are old enough to get vaccinated without parental consent (18-21-year-olds), one five are vaccine resistant. 

Our survey suggests that awareness of the J&J pause was extremely high. Despite this, vaccine hesitancy/resistance did not increase for responses received after the pause, and our analysis of repeat respondents suggests a small but systematic shift decrease, largely because a fair number of people who were vaccine hesitant in early April were vaccinated by late April. In short, it seems very unlikely that the pause had major negative effects on vaccine attitudes. Our confidence in this finding is increased by the observation that we see these effects using two different kinds of analysis: comparing over-time changes across different subjects and, separately, within the same subjects interviewed before and after the pause. Still, it is highly likely that vaccine rates will continue to decelerate over the next few months as those who intend to get the vaccine as soon as they can are able to do so. 

This report is an update of our February report on attitudes and vaccination rates of healthcare workers. The essential patterns are similar, with the obvious exception being the increased rates of vaccination. We focus on vaccination, vaccine resistance, vaccine hesitancy, vaccine access, and vaccine refusal. These are measured as follows: 

Vaccination status is based on self-reports that someone has “already” been vaccinated. Note that we do not ask whether someone has received one or two doses of a vaccine, or which vaccine they have received. 

Vaccine resistance is the proportion of individuals in a given category who indicate that they “would not get the COVID-19 vaccine” if/when it is available to them. 

Vaccine hesitancy is defined as preferring to get the vaccine “after at least some people I know” or “after most people I know.” 

Vaccine access considers those in one of the high priority groups—individuals 65 and older or healthcare workers—who report having the vaccine available to them. This is defined as anyone in this group who says they have already been vaccinated or have a vaccine available to them (regardless of whether they have been vaccinated). 

Vaccine refusal is when an individual in a high priority group reports having the vaccine available to them, but that they will not get vaccinated. 

Vaccination rates are higher among healthcare workers than among non-healthcare workers;1 however, vaccine resistance (24% for healthcare workers versus 21% for non-healthcare workers) and hesitancy rates (29% versus 31%) are nearly identical (see Figure 1). 

Vaccine resistance has remained level since January for both healthcare workers and non-healthcare workers; however, hesitancy has dropped substantially for both groups since January (from 37% to 29% among healthcare workers, and from 41% to 31% for non-healthcare workers). This suggests that as the vaccine campaign moves forward, as one would expect, there has been a gradual shift from those who are vaccine hesitant to being willing to be vaccinated. 

● Parents are more vaccine hesitant and resistant than non-parents (in terms of willingness to vaccinate themselves) across all socioeconomic and demographic groups we compared. However, this pattern is largely driven by younger mothers, who are far more vaccine resistant than younger women who are not mothers. Older parents and fathers show little difference from their non-parent peers. 

● Parents’ willingness to vaccinate their children closely matches their willingness to get vaccinated themselves. Mothers are far more reluctant than fathers to do both. 

● Parents without a 4-year college degree are far more likely to be vaccine hesitant and resistant than their counterparts with a 4-year college degree or higher. Among the more highly educated, parents and non-parents hold similar views. 

● Parents who earn less than $75,000 per year are far more reluctant to get vaccinated than their non-parents counterparts. The gap is far smaller in households earning over $75,000 per year. 

● Parents of all races are more reluctant than non-parents to get vaccinated themselves, but this gap is largest among African American parents, nearly three quarters of whom are vaccine hesitant or resistant. 

● We see similar gaps between parents and non-parents across party and regional lines, and between urban, suburban, and rural residents. 

● Parents are more vaccine hesitant and resistant than non-parents (in terms of willingness to vaccinate themselves) across all socioeconomic and demographic groups we compared. However, this pattern is largely driven by younger mothers, who are far more vaccine resistant than younger women who are not mothers. Older parents and fathers show little difference from their non-parent peers. 

● Parents’ willingness to vaccinate their children closely matches their willingness to get vaccinated themselves. Mothers are far more reluctant than fathers to do both. 

● Parents without a 4-year college degree are far more likely to be vaccine hesitant and resistant than their counterparts with a 4-year college degree or higher. Among the more highly educated, parents and non-parents hold similar views. 

● Parents who earn less than $75,000 per year are far more reluctant to get vaccinated than their non-parents counterparts. The gap is far smaller in households earning over $75,000 per year. 

● Parents of all races are more reluctant than non-parents to get vaccinated themselves, but this gap is largest among African American parents, nearly three quarters of whom are vaccine hesitant or resistant. 

● We see similar gaps between parents and non-parents across party and regional lines, and between urban, suburban, and rural residents. 

In this report, we analyze the trajectory of the pandemic in Massachusetts, from late April 2020 to March 1, 2021, examining public health behaviors and approval of Governor Charlie Baker’s handling of the crisis. For information on other states, please consult our interactive dashboard. 

● Education is strongly correlated with vaccination rates (29% of respondents with graduate degrees vs 9% of those with a high school education or less) and vaccine resistance (8% vs 30%). 

● Higher income respondents are far more likely to be vaccinated than lower income (24% for highest income cohort vs 9% for lowest income cohort); and have much lower levels of vaccine resistance (11% vs 29%). 

● Asian Americans (17%) and Whites (16%) have the highest vaccination rates, and African Americans (12%) and Hispanics (9%) the lowest. Asian Americans have by far the lowest levels of vaccine resistance (10%), with Whites (21%), Hispanics (22%), and African Americans (26%) all at far higher levels. 

● Democrats and Republicans have similar vaccination rates (17% vs 16%), but only 9% of Independents have been vaccinated. Independents (31%) and Republicans (30%) have far higher levels of vaccine resistance than Democrats (11%). 

● Women are less likely to be vaccinated than men (13% vs 17%), and more likely to be vaccine resistant (25% vs 18%). 

● Rural areas have lower vaccination rates (12%) than urban and suburban (15% each), and far higher vaccination resistance (29% for rural areas; 22% for suburban; 16% for urban). 

● There are substantial variations of vaccine resistance by state: Massachusetts with the lowest levels (9%) and Oklahoma and North Dakota with the highest (33%). 

‍

The major pattern we find in the data is a substantial relaxing of social distancing behaviors since the spring of 2020, with some tightening in the late fall/early winter that year. In Figure 1, we show trends in 8 different activities outside of the home that people reported engaging in within the 24 hours prior to taking our survey. The activities include going to work, to the gym, to a bar or café, to a place of worship, to the doctor, visiting a friend, taking mass transit, or being in a room with people who live outside one’s household. 

Residents of New Jersey closely followed the social distancing guidelines in the spring of 2020, but progressively relaxed their adherence to health recommendations over the summer and early fall of 2020. People once again became more careful about social distancing as COVID-19 case counts increased towards the end of 2020. Our most recent data indicate that as case numbers decreased in 2021, behaviors relaxed as well. 

In Figure 1, we show trends in 8 different activities outside of the home that people reported engaging in within the 24 hours prior to taking our survey. The activities include going to work, to the gym, to a bar or café, to a place of worship, to the doctor, visiting a friend, taking mass transit, or being in a room with people who live outside one’s household. 

The data show a notable uptick in all of these behaviors from April to mid-November. We observed a similar trajectory for each of the eight activities: low points in spring 2020, followed by high points in the fall, declines in the early winter, and finally a recent increase in February 2021. For instance, there was a large surge in people going to work outside of their home from late April to mid-November 2020 (19% to 40%), a subsequent drop to 31% in January 2021, and then a jump to 39% in February 2021. 

The percentage of New Jersey residents who reported having been in a room with people other than members of their household more than doubled from late April to mid-November 2020 (21% to 45%) and was at 36% in February 2021. Reports of most other indoor activities also increased from late April to mid-November, although less dramatically. The percentage of people who reported going to a cafe, bar, or restaurant jumped from 5% to approximately 12%, the percentage that indicated going to a doctor or hospital increased from 7% to 11%, and the percentage taking mass transit rose from 1% to 7%. Visiting a friend represents a minor exception: such reports increased between April and late August, decreased until mid-November, and then peaked in early December. 

In this report, we examine the demographic correlates of vaccination, vaccine hesitancy, and vaccine resistance among the subset of 1,797 respondents in the COVID states survey who indicated that they are healthcare workers. We find that education, income, gender, race/ethnicity, and partisanship are strong predictors of vaccination rates recorded thus far, as well as of vaccine hesitancy and resistance. 

Healthcare workers represent a microcosm of US society, and a consequential one. They are polarized in terms of income and education, ranging from physicians with extensive postgraduate education, to cleaning staff with less than a high school education. Healthcare workers also have taken a disproportionate brunt of the national response to COVID-19, in terms of the physical and emotional labor to care for millions of sick people; to bid goodbye to hundreds of thousands; and to confront, by the nature of their work, the increased risks of exposure to the virus. 

The 2020 election produced two distinct perceived realities for the United States public. The first perceived reality holds that the election was conducted fairly, and that Joe Biden won. Individuals who hold this vision of reality feel their votes were counted accurately, and that the events of January 6th were repugnant. This is also the version of reality for 60-70% of the US population. This reality is documented by court cases, our government officials (Republican and Democratic) in charge of administering the election, and the credible news media. 

The second vision of reality holds that the election was essentially corrupt—driven, in particular, by illegal mail-in ballots, noncitizen voting, and voting machine fraud. In this perceived reality, Biden is a usurper to the presidency. Further, the people who stormed the Capitol had a point, although the worst actions were taken by individuals affiliated with Antifa activists masquerading as Trump supporters. This is the expressed reality of former President Trump, certain Republican leaders, a subset of conservative media, and certain corners of the social media ecosystem. This is also the reality for 25-30% of the US population, and roughly half of Republicans. 

These two perceived realities are not created equal. The facts, as presented by the fact sorting institutions of the government and media, unambiguously support the first. However, perceptions—especially when they are shared by a sizable number of people and organized in a way that provides access to power—have consequences. The data below suggest that these two perceived realities will continue to define US politics, at least for the near future, because they are clustered around the two parties, and in each perceived reality, the other party is not just an opponent, but an enemy of democracy. In each case, the logical conclusion is that the assumption of power by the other party represents a potentially permanent surrendering of power. Below, we summarize public opinion regarding election integrity and the events of January 6th, and close by discussing the implications of these findings for American politics over the next 4 years. 

In our December 2020 survey wave, we included a series of questions aimed at assessing the state of opinion regarding these topics. Some key findings are: 

● Two thirds of respondents are at least somewhat concerned about the quality of the education they (if they are students) or their children (if they are parents) are receiving during the pandemic. This is true across partisan, racial, and income groups. 

● A majority of parents of K-12 students indicate that their children learned less than they would have without the pandemic, including pluralities of primary school (grades K-5) parents and majorities of parents of middle (grades 6-8) and high (grades 9-12) school-age children. This also includes, at minimum, substantial pluralities, and in most instances, majorities, across all demographic groups across all grade levels. 

● We find some evidence of a socio-economic divide, with wealthy Americans more likely than other income groups to believe their high school-age children are learning less during the pandemic than other income groups. But this pattern does not extend to younger children. 

Gun purchases in the United States have increased dramatically since the start of the COVID-19 pandemic. Based on FBI data, the National Shooting Sports Foundation reported 21 million background checks for firearms sales in 2020, up from 13.2 million the previous year and surpassing the previous record of 15.7 million in 2016. The biggest spikes occurred in March and June, as shown in Figure 1. While March is typically a month of high gun buying, this year saw a much sharper surge. June’s spike in gun purchases is even more abnormal compared to previous years when gun purchases are typically close to their low point. 

Our survey helps in understanding this increase. We find that gun purchases were particularly likely among attendees of protests against racism or police violence. These protests reached massive proportions during the first week of June, the second-highest week ever for background checks at the time. However, it is doubtful that protesters themselves—about 5% of the adult population by our estimates—can fully account for this increase. 

Others may have purchased guns in reaction to this widespread unrest, as well as the sweeping government-mandated closures that coincided with the first peak in March. First-time gun purchases were more likely among African Americans, while existing gun owners were significantly more likely to buy guns if they reported not wearing masks or if they had contracted COVID-19. The most common reasons for purchase, by far, were protection against crime (70%) and target shooting or hunting (47%), though over a quarter (32%) said they were reacting to lockdowns, fears of the government, COVID-19, or the 2020 election. 

We evaluate the results of two experiments testing different communication strategies to increase people’s willingness to take the COVID-19 vaccine. Key findings are that a variety of messages are effective for reducing vaccine hesitancy and resistance, while messengers are less effective and − in the case of partisan messengers − have the potential to increase resistance among political out-groups. The most effective strategies evoked scientists and people’s own doctors as endorsers of vaccination. We found, also, that messages evoking harm reduction and “people you know” were more effective in counties where the virus is spreading more quickly. 

We present a variety of descriptive results on attitudes and beliefs regarding COVID-19 vaccines. We find a notable consensus across population subgroups regarding order of prioritization for receiving vaccines, which roughly aligns with current policies. We additionally find that while there are generally positive attitudes and beliefs regarding vaccines, there are also some negative attitudes and misbeliefs that are especially common within certain demographics (political independents and, to a lesser extent, Republicans; Black respondents; and individuals ages 25-44). 

The reported results are based on data collected from 24,682 people between December 16, 2020 and January 10, 2021. For these descriptive analyses, we apply survey weights using national benchmarks for race, gender, age, education, Census region, and urbanicity. 

Public health officials have expressed concern that multi-household holiday gatherings might have driven a January surge of COVID-19 cases. As noted in report 31, social distancing behaviors have generally been improving since late October, including in our most recent wave conducted over the holidays. However, the scope of our data collection regarding social distancing, which focused on proximity to others in the 24 hours preceding survey response, might have missed important spikes in proximity around Thanksgiving, Christmas and New Year. Therefore, here we evaluate the responses from a survey conducted between December 16th and January 10th, to a separate question asking whether people had gathered for the holidays in November and December. 

1) Average wait times for COVID-19 test results continue to fall steadily (from 3.9 days in March to 2.2 days in December). 

2) Racial disparities still exist in testing times: in December, Hispanic and Black respondents waited 0.7 and 0.4 days longer than white respondents, respectively. 

3) In December, the average respondent waited 2.4 days between seeking a test and obtaining a test. They waited another 2.2 days to receive test results. 

4) Testing is still too slow to support effective contact tracing. 

In report 25, published in November, we evaluated public support for taking more aggressive measures to limit the spread of COVID-19. The key finding is that there is broad support for taking more steps to limit the spread of the disease. In a just completed survey, we repeated these questions, finding that opinions are virtually unchanged since November: large majorities across the 7 issue areas support taking more aggressive steps (see Figure 1). The measure with the least support, which would be the most drastic step − requiring most businesses other than grocery stores and pharmacies to close − has exactly the same level of support we observed in November: 60%. The only step that has seen statistically significant erosion in support is restricting international travel, which has edged downward from November, from 88% to a still quite high 84%. 

Since April 2020, the COVID States Project has been surveying respondents regarding social distancing behaviors and adherence to health recommendations more generally. Our November report, looking at the data from April to November, with a partial exception of mask wearing, found a steady relaxation of these behaviors over time. In this report, we update our November results, finding that this trend has partially reversed, perhaps due to the surge of COVID-19 cases in the last two months. Mask wearing reached its all-time high in our December/January survey, and social distancing behaviors have substantially increased since October (though still far from April levels). 

● Parents with children at home are more likely than respondents without children at home to have experienced all five economic hardships, while women with kids at home are more likely than men with kids at home to have experienced four of the five hardships. The overall gaps range from a low of 4 percentage points for eviction from their home (6% for parents vs. 2% for non-parents), to highs of 11 points for taking pay cuts (26% vs. 17%) and being unable to make a rent or mortgage payment (21% vs. 10%) 

● Economic hardships are fairly evenly distributed across partisan groups and regions of the country. 

● Respondents who report having had COVID-19 (whether diagnosed by a medical professional or otherwise) are more likely than respondents who report not having had had COVID-19 to have suffered a series of five economic hardships we included in the survey, including losing a job, being evicted from their home, being unable to pay rent or their mortgage, taking a pay cut, or having to cut back on work to take care of a child. 

● Hispanics are more likely than other racial/ethnic groups to have experienced all five economic hardships. 

● We find fairly wide variations across the 50 states in the percentages of residents having experienced economic hardships, and no consistent geographic patterns. COVID-related hardships appear to be generally widespread across the country. 

• While 67% of Americans say President-Elect Biden won the election, 17% suggest that Donald Trump is the winner. Thirty-nine percent of Republicans and 3% of Democrats reported thinking Trump is probably winning or definitely winning. 

• Overall, 38% of Americans lack confidence in the fairness of the 2020 presidential election. That number is especially high among Republicans (64%) and Trump voters (69%) compared to Democrats (11%) and Biden voters (8%). 

• There are large partisan gaps (over 40 percentage points) in public concerns about mail-in fraud (85% of Republicans, 38% of Democrats), inaccurate or biased vote counts (84% of Republicans, 44% of Democrats), and illegal votes from non-citizens (81% of Republicans, 34% of Democrats). 

• Asked about acceptable reactions to an unfair election, 45% of Americans approved of protesting on social media, 38% of protesting in person, 18% approved of violating laws without violence, and 8% of using violence. Non-violent law breaking was approved by 23% Democrats and 17% Republicans, violence by 10% Democrats and 8% Republicans. 

• 69% of Americans trust the Supreme Court to handle the election, 43% trust the news media, 31% trust social media companies. Trump is trusted by 39%, Biden by 59%. 

In November, we find quite high levels of support for all seven restrictive measures, ranging from lows of 63% of Massachusetts residents supporting closing most businesses and 71% supporting prohibiting in-person teaching in K-12 schools, to highs of 89% and 88% supporting restricting international travel to the US and imposing stay-at-home requirements, respectively (Figure 1). Between the extremes, 76% support restricting domestic US travel, 79% support restricting restaurants to take-out only, and 82% support canceling major sports and entertainment events. These numbers are generally slightly higher than national support for these measures (identical for restricting international travel, and up to six points higher in every other case).1 

The percentages varied only modestly from October levels, with the largest change being a 3-percentage point decline in support for closing businesses. For all other measures, the difference from October to November is 2 points or less. 

The number of COVID-19 cases in Illinois has risen sharply over the last two months, from about 2,000 new cases a day in late September to 10,000 or so currently. In this report we evaluate whether there have been significant changes in the behaviors that facilitate the spread of the disease. The COVID States Project has been conducting a 50-state survey roughly once a month since April about attitudes and behaviors around COVID-19. Here we focus on public-health related behaviors that facilitate/inhibit the spread of the disease. It has become apparent that behaviors of particular importance are (1) those which bring people into indoor proximity; and (2) mask wearing. 

COVID-19 is transmitted from human to human, and as such, the key behavior that drives the spread of COVID-19 is close proximity among people. In particular, large groups of people indoors facilitate the occurrence of super-spreading events, where a single contagious individual may infect multiple people at once. Recent research highlights, in particular, the role of restaurants and gyms in driving the early spread of COVID-19. Figure 2 presents the trends for a set of activities that bring people from different households into indoor spaces. We note the general upward trend in many of these activities since April, with reports of having been in a room with people who are not members of the household in the preceding 24 hours jumping from 26% in April to 45% in October. Large group activities have particularly jumped in frequency. Reports of being in groups of 11 to 100 or more in the preceding 24 hours more than doubled, from 2.4% of respondents in April to 6.4% in October (Figure 3). 

(1) Substantial majorities (6 in 10, or more) of Americans support all seven restrictive measures. 

(2) Majorities of Americans -- and in most cases large majorities -- in every state support five of the seven restrictive measures. The exceptions are prohibiting K-12 schools from teaching in-person classes and closing most businesses other than grocery stores and pharmacies. In the latter two cases, support falls below 50% in three states for businesses and two states for schools (yet both are still supported by majorities in all other states). 

(3) Support is particularly strong for prohibiting international travel and requesting that people stay at home and avoid large gatherings. 

(4) Americans across nearly all subgroups (partisanship, race, gender, age) and states are, in general, least supportive of closing businesses other than grocery stores and pharmacies, followed by prohibiting K-12 schools from teaching in-person. 

(5) The greater a respondent’s concern about contracting COVID-19 for either their family members or themselves, the greater their support for restrictive measures. Similarly, belief that COVID-19 is increasing in a respondent’s state is also positively related to their support for restrictions. 

(6) Families with children at home are about as likely as other respondents to support restrictive measures, including closing K-12 schools to in-person classes. 

(7) Americans are starkly divided along partisan lines. Democrats and Biden voters are substantially more supportive of nearly all of the queried measures than Republicans or Trump voters, especially for closing in-person K-12 schools, non-essential businesses, and eat-in restaurants. Nevertheless, all four subgroups (Democrats, Republicans, Trump Voters, Biden Voters) are least supportive of closing non-essential businesses. 

The critical question we address in this report is how have behaviors changed during the pandemic? Our data confirm a substantial relaxation of many of the behaviors that helped slow the spread of the disease in the spring (see Figures 1 and 2). For example, take restaurants: in April and May, in Massachusetts, fewer than 5% of respondents reported going to a restaurant in the preceding 24 hours. This number rose to 15% in October. For other kinds of proximity to non-household members in enclosed spaces, the number of people taking part doubled from 22% in April to 45% in October—i.e., twice as many respondents said they had been close to non-household members. Going to a gym jumped from less than 1% to 7%. These patterns are not just the result of individual choices, but of policy decisions, since many of these establishments were closed in the spring. 

The next US President will confront impacts far broader than COVID-19 infection alone in trying to deal with the pandemic. Neither candidate provided much detail about their response to the mental health consequences of the disease and efforts to contain it. In line with our May results, our survey indicates that the next administration will lead a country where unprecedented numbers of younger individuals are experiencing depression, anxiety, and, for some, thoughts of suicide. These symptoms are not concentrated among any particular subgroup or region in our survey; they are elevated in every group we examined. Our survey results also strongly suggest that those with direct economic and property losses resulting from COVID-19 appear to be at particular risk, so strategies focusing on these individuals may be critical. 

Despite the headwinds confronting the country in facing COVID-19, approval of most governors’ and President Trump’s COVID-19 response did not change greatly between September and October in our survey. President Trump’s approval has ticked up a point to 35%, which is still problematically low approval for the most salient issue in the 2020 election by far (in 49 out of 50 states, Alaska being the one exception, COVID-19 is ranked as ”the most important problem”). More generally, after suffering a decline in the spring, the President’s approval rating has been flatline around his October rating of 35%. (For all approval ratings for the President and all of the governors, see the tables in the appendix.) To compare, governors, as a group, have seen a gradual decline in their approval ratings between late April and September; a 1 point increase from September (47%) to October (48%) marks the first time they have improved, or even held steady since our initial survey wave. Despite this, approval of governors is generally higher than that of the President. In 42 states, the governor has a higher approval level than the President, and in 2 states the levels are the same; in 6 states, the President has a higher approval level, but in no state does the approval of the President exceed that of the governor by more than the margin of error for the survey in that state.1 Per-state approval of the President roughly tracks partisanship of the state, and the only state where his approval is above 50% is Alabama. In most of the 10 swing states, the President hovers around 35%, from a low of 29% in New Hampshire to a high of 42% in Ohio. Alaska and Iowa stand out as states that he won handily in 2016, but now finds very low approval -- 34% and 32%, respectively. Other states he won in 2016 and now lingers at the national average or below are Michigan and Wisconsin. Pennsylvania --perhaps the most pivotal state in the 2020 election -- runs against the trend somewhat, showing approval ratings below the national level in April, and now several points above (at 38%). 

Perhaps unsurprisingly, the most highly cited issue is COVID-19, mentioned as the top problem facing the nation by nearly one third (32%) of all respondents nationally. While the percentages vary widely, from a low of 13% in Wyoming and Alaska, to highs of 44% and 42% in the District of Columbia and Hawaii, respectively, COVID-19 was the number one problem mentioned by respondents in every state except Alaska, where climate change comes out on top. (In Alaska, 16% named climate change as the number one problem facing the nation.) Figure 1 presents the percentages of respondents naming COVID as the number one problem facing the nation in each state. 

Other problems mentioned by at least 5% of respondents include racism (10%), the economy (8%), healthcare (7%), and crime and violence (6%). These topline figures suggest that the issues both candidates have sought to emphasize are indeed among the most frequently cited problems. Yet, they also suggest that the campaign, in the aggregate, is being fought out more squarely on turf preferred by Biden. Biden's preferred issues rank 1st (COVID-19), 2nd (racism), and 4th (health care) among the public; in contrast, Trump's preferred issues rank 3rd (the economy) and 5th (crime and violence). 

In every month, April through October of 2020, we surveyed individuals in every state about how federal and state governments are reacting to the pandemic. We found a remarkably consistent picture of public opinion: respondents prefer state governments over the federal government when it comes to COVID-19. Out of 8 waves in 50 states & DC − a total of 408 surveys at the state level − in 402 state-level surveys more people in the state felt the state government was reacting “about right” to the COVID-19 outbreak as compared to the federal government; and only 6 times did people in a state choose the federal government over their state government. 

The COVID-19 pandemic has been the defining issue of the 2020 general election, and the major party candidates have sent markedly different signals regarding its severity, how the federal government should respond, and which behaviors individuals should and should not be engaging in while the virus continues to spread. As such, it is perhaps unsurprising that there is a small but significant amount of movement between candidates − on net, away from Donald Trump − associated with levels of concern regarding the pandemic and perceptions regarding the federal government’s response. 

Of course, it is important to note that despite the patterns we observe here, the vast majority of voters’ behavior has remained stable over time, and not all of the changes we observe are attributable to the pandemic. However, these results do suggest that, when provided with a strong signal regarding significant material differences between the major party candidates on an extremely salient issue, some voters are willing to update their candidate preferences accordingly. With days remaining until election day, and with the number of COVID cases surging, there is likely little the president can do to change voters’ evaluations of his pandemic response. 

‍

COVID-19 has been a massive disruption in 2020, exacerbated by the spread of misinformation on social media. In this report we ask who is sharing COVID-19 fake news on Twitter, and which URLs are being shared? We show that older people are more likely to share URLs from fake news domains, despite being less likely to believe COVID-19 misinformation. We also show that Republicans are more likely to share fake news than Democrats and Independents. Finally, the fake news domain that has been shared the most in COVID-19 tweets is the Gateway Pundit, which, throughout the COVID-19 pandemic, has received an order of magnitude more shares than even the second most popular fake news domain. 

● Average testing times have fallen since April (from 4.0 days to 2.7 days in September) 

● Across all months, Black respondents wait almost an entire day more than white respondents for their test results (4.4 days to 3.5 days, on average) 

● In our September and August waves, the average respondent waited 6.2 days between seeking a test and receiving test results 

● Only 56% of those who tested positive for COVID-19 report being contacted as part of a contact tracing attempt 

The above analysis highlights the potential for the popular vote and the electoral college to shift toward Biden after election night. Trump will have a substantial lead among those who vote the day of the election; and Biden an even large margin lead those who vote by mail. How much this translates into a post-election-day blue shift depends on how quickly states count votes by mail. There are 9 competitive states that, because they either allow votes to arrive after election day or do not start counting until election day, will plausibly lag substantially in counting votes. This possibility puts a particular onus on the media and on political elites: For the media, to set expectations that vote totals may shift substantially after election day, very likely toward Biden; and for political elites, to make sure the voting systems are well resourced and to not exploit the time lag in reporting results to undermine the legitimacy of the election. 

To date, Congress has passed four COVID-19 relief packages totaling about $3 trillion, the most recent of which, the CARES Act, was passed on March 27th, 2020. The House of Representatives subsequently passed the $3 trillion Heroes Act on May 15th. The Senate has yet to take up the legislation or pass a fifth bill of its own. Throughout the summer, the House, Senate, and White House have engaged in on-again, off-again talks aimed at agreeing on a fifth relief package. The Senate has consistently favored a smaller bill, ranging from $500 billion to $1 trillion. Most recently, the House introduced a revised bill valued at about $2.2 trillion. The White House, represented by Steven Mnuchin, along with the Speaker of the House, Nancy Pelosi, have resumed negotiations over the shape of the package. 

While both sides claim to support such a package, they disagree not only on the amount, but also on the targets for this funding. Democrats favor a combination of direct payments to Americans, extended unemployment insurance, and aid to hospitals, schools, small businesses, the Post Office, and state and local governments. Republicans oppose aid to state and local governments and favor more limited unemployment insurance benefits. 

But what does the public think? We surveyed 20,315 respondents between September 4-27, 2020 on attitudes regarding the next COVID-19 relief bill. We asked respondents if they supported such a bill, as well as the types of relief they believed it should include. We also probed whether or not they had received the $300-$400 supplemental unemployment benefit authorized by President Trump’s executive order of August 5th. 

Scholars and public health officials have expressed growing alarm over what some have termed a “misinfodemic” − a parallel epidemic of misinformation − around COVID-19. Indeed, conspiracy theories, from the Plandemic pseudo-documentary to QAnon, fuel rising skepticism about scientific facts across many areas of public life, and in recent months especially with respect to COVID-19. Misperceptions, which can rapidly spread from obscurity to mass exposure via social media, may have the capacity to hinder the efficacy of public health efforts aimed at slowing the spread of the pandemic. Especially concerning, encountering false claims online may ultimately reduce the willingness of some Americans to get a COVID-19 vaccine when it becomes available. 

In this report, we assess respondents’ acceptance of 11 false claims that have circulated online since the beginning of the pandemic. The statements we use include six false claims about conspiracies or risk factors and five false purported preventive treatments for COVID-19. For the conspiracies/risk factors, we asked respondents whether or not they thought each claim was accurate, or whether they were unsure about its accuracy. For the false preventive treatments, we asked participants whether or not they believed the purported treatment was effective, or whether they were unsure about its efficacy. 

Controversy has erupted over the possible release of a COVID-19 vaccine prior to the November 3rd election, with President Trump hinting that a vaccine may be ready by October, while the CDC has instructed states to be prepared to distribute a vaccine at that time. These assertions have prompted a backlash from some public health experts, countering that this timeline is extremely improbable. 

Concerns over the possible politicization of emergency use authorization by the Food and Drug Administration (FDA) of a COVID-19 vaccine prior to the completion of phase 3 trials, in turn, has prompted the various pharmaceutical companies currently working on vaccines to prepare an unprecedented joint public statement that they will not submit a vaccine to the FDA for approval until they have compelling scientific proof that it is safe and effective. The political debate occurring against the backdrop of a presidential election raises concerns that many Americans may be unwilling to accept a vaccine because they do not trust the government or its agencies to prioritize safety and efficacy over politics. 

Meanwhile, Democratic vice presidential candidate Kamala Harris commented during a CNN interview on getting a COVID-19 vaccine prior to the election: “I would not trust Donald Trump.” She subsequently amended her statement, saying in part, "I would trust a vaccine if the public health professionals and the scientists told us that we can trust it." In this report, we explore the issue of public trust regarding the handling of COVID-19 across 15 institutions and individual leaders and its link to public willingness to be vaccinated if a vaccine becomes available. 

The August wave of our COVID States survey indicates a continued gradual decline in approval of governors’ management of COVID-19, and a small improvement in President Trump’s management of COVID-19. The average governor saw their approval slip another 3 points in the last month, to 48% approval (versus 63% in late April). President Trump, who started from a much lower point, rebounded slightly from his low point of approval of 32% in July to 34% in August. See Figure 1 and Table 1 for trends in approval for governors by state, and Figure 1 and Table 2 for trends in approval for President Trump. The governors of Maryland, Massachusetts, and Connecticut are all at 65% approval or above. States with governors that have notably low approval ratings (below 40%) include Arizona, Florida, Georgia, Iowa, Hawaii, Idaho, Mississippi, Oregon, Oklahoma, Missouri, South Carolina, and Texas.

‍Governors who have suffered particularly precipitous drops in the last month include the governors of California (Newsom), Virginia (Northam), and Florida (DeSantis). Newsom dropped from 58% approval to 47%, Northam from 59% to 46%, and DeSantis from an already low 40% to 29%. Notably, despite the general stability in presidential approval, Trump’s approval in Florida of his management of COVID-19 also dropped sharply, from 39% to 29%.

‍Approval for governors dropped since April among their co- and anti-partisans; but generally, more so among anti-partisans. Average approval for a Democratic governor among Democrats dropped 8 points since April; and 16 points among Republicans. Approval for Republican governors dropped 12 points among Republicans; and 19 points among Democrats. See Figure 2 for partisan patterns in approval by state, and Figures 3 and 4 for averaged approval for Republican and Democratic governors by respondent’s party affiliation. (Note that these subsamples are smaller, and thus the trends are noisier at the state level.) 

‍

An increasing number of school districts across the United States have announced that they will provide entirely remote learning when school resumes this fall. Others, like New York City, continue to consider alternatives, including hybrid models in which students attend school part-time. While many other countries have succeeded in reopening schools without a resurgence of cases (with some notable exceptions), they reopened in a very different context, with rates of infection in the community far lower than in many places in the United States. On the other hand, the federal government, and some state governments, continue to focus on the importance of reopening. Another complication is the role of teachers’ unions, who have expressed discomfort about their teachers’ safety and their ability to provide a safe environment for students. And among schools already reopening, some students have already tested positive for COVID-19. 

Parents of school-age children thus confront difficult decisions about their children’s education - decisions that will hinge on whether they believe schools can be made safe. We surveyed 19,058 adults in the 50 U.S. states and the District of Columbia for 2 weeks beginning July 10 and ending July 26, asking about their perception of the safety of returning to school. We also asked a subset of respondents whether they support reopening schools for in-person classes in the fall. We further aimed to understand whether these comfort levels vary depending on rates of infection in their community, and on other attitudes about COVID-19. 

● The protests were massive and truly national in scope. Even the states with the lowest protest participation rates, such as North Dakota, West Virginia, and Wyoming, had about 2% of adults report participating. 

● The rate of participation was strongly negatively correlated with Trump vote share in 2016. We show a scatterplot of reported protest participation rate against Trump vote share in Figure 2. 

step before seeking approval from the U.S. Food and Drug Administration (FDA). According to NIAID director Anthony Fauci, vaccines may be widely available in the U.S. by spring 2021 if these trials are successful. 

But should these vaccines become available, will Americans accept them? Between July 10 and July 26, we surveyed 19,058 adults in all 50 U.S. states and the District of Columbia. We asked about the likelihood that they would seek vaccination for themselves, and for their children. We also asked about the factors that would influence their decision making. 

We find that, overall, 66% of adults would be somewhat or extremely likely to vaccinate themselves; 66% would be somewhat or extremely likely to vaccinate their children. These rates vary markedly between states, as shown on the figure below. 

Rapid turnaround of testing for COVID-19 infection is essential to containing the pandemic. Ideally, test results would be available the same day. Our findings indicate that the United States is not currently performing testing with nearly enough speed. In our large (19,058 respondents) national survey, conducted between July 10 and 26, we asked whether respondents had been tested for COVID-19 and how long they had waited to get results. Our finding: 37% of those who had been tested by nasal swab received results within 2 days, and the average wait time was 4.1 days; with 31% of tests taking more than 4 days, and 10% 10 days or more. Further, there are few signs that turnaround times are diminishing. For individuals who responded that their last test had been in April, they had waited on average 4.2 days to get results; and for individuals tested in July, 4.1 days. 

Overall, support for voting by mail is quite high, with 64% of likely voters in our survey1 supporting making it easier to vote by mail, and only 23% opposing. 

Similarly, more Americans who say they plan to vote in 2020 support than oppose universal extension of vote by mail. A total of 66% of likely voters report that they would support giving every American the right to vote by mail in November, with 22% opposing. 

The latest update to our executive approval series reveals a mix of continuity and change. The most obvious difference in the circumstances confronting the nation’s executives in the latest wave (July 10-26, 2020, N=19,052) compared to the prior wave (June 12-28, 2020) is that far more states are seeing increases in COVID cases than in early June. For instance, in the Northeast in our late June wave, all but one state (New Jersey) were trending downward in COVID cases relative to our early June wave. In late July, all but two states in the region are trending upwards in COVID cases. (New Jersey, ironically, represents one of the two improving states, with the second being Maine). Despite these worsening trends, many governors in the Northeast (excepting Maine, where Democratic Governor Mill’s approval rating stands at a middling 54%) remain quite popular, and have seen their approval ratings remain relatively stable (e.g. Massachusetts and Vermont) or increase (e.g., Rhode Island and Maine). Two governors in the region — Governor Murphy (D) of New Jersey and Governor Sununu (R) of New Hampshire — have faced substantial erosion in public support for their management of the pandemic since June, by 6 and 8 percentage points, respectively. 

The initial response to a crisis typically depends on the executive branch of government, because they may act more rapidly than legislative and judicial branches. For COVID-19 in particular, the focal decision-makers have been the president and the governors of the 50 states. In the eyes of the public, how have the president and governors responded? Figure 1 plots the change in approval since April for presidential and gubernatorial responses to COVID-19 for all 50 states plus the District of Columbia. 

We find a broad, nearly nationwide pattern of declining executive support. Figure 1 reveals several patterns1. First, the average governor has experienced a 10-point decline in approval from late April to late June, and in only 5 states—Hawaii, Michigan, New Jersey, South Dakota and Vermont*—have governors’ approval ratings increased since late April. In the remaining states, approval has declined. Second, the President began (at the outset of our series in late April) with a relatively low approval rating, below that of the governor in all 50 states, and has declined further—by 8 points on average. Only one governor—Governor Doug Ducey of Arizona— now has an approval rating in their state lower than that of the president. 

Even as all 50 states continue to move towards full reopening, Americans remain more cautious than their governors. Only 15% of survey respondents say that the country should be reopened for business right away, a 4 percentage point increase from the response to that question in early May. More than half of those polled (57%) are in favor of waiting at least a month or more before resuming business activity. 

Partisan differences in attitudes towards reopening continue to grow, with 30% of Republicans now in favor of immediate reopening compared to 4% of Democrats. That represents an increase of over 20 percentage points for Republicans over the last month, compared to only a 2 percentage point increase for Democrats during the same period. 

The residents of six states — Massachusetts, New York, Rhode Island*, New Jersey, New Hampshire, and Maryland — are especially reluctant to reopen right away, with 10% or fewer of respondents in those states in favor of immediate reopening. At the other end of the spectrum, a quarter or more of the respondents in Idaho*, Montana*, Alaska* and Wyoming* say that the country should reopen immediately. The states with largest shifts in attitude towards immediate reopening are Alaska*, Wyoming*, South Dakota*, Oklahoma, Ohio, and New Mexico*, with an increase in support of 15 percentage points or more over the last month. 

‍

A strong majority of Americans support making it easier to vote by mail (VBM). Three-in-five (60%) Americans support making it easier to vote by mail this November, while only 16% oppose (24% neither support nor oppose). Majorities of Americans in all but four states -- Arkansas, Mississippi, Tennessee, and Wyoming* -- support making VBM easier. At the other extreme, over 70% of Americans in four states (Oregon, Washington, California, Colorado) and the District of Columbia*, support making VBM easier. 

Democrats and residents of battleground states are more likely to support making mail ballots easier. Overall, 80% of Democrats support making it easier to vote by mail, 5% oppose; for Republicans, 45% support and 32% oppose. Independents, by a large margin, also support greater access to mail balloting (56% support, 14% oppose). Americans in red states were less likely to support an easier vote by mail in November (53% support, 20% oppose) compared to those living in 

Americans are in less of a hurry to reopen than their governors. Despite the movement toward economic reopening in all 50 states, we see little evidence that Americans are clamoring to speed up the reopening process. Overall, we found only a five percentage point increase from our prior survey wave in late April in the percentage of Americans preferring that the economy reopen “immediately.” A sizable majority (60%) of Americans continue to prefer that the country wait at least four more weeks prior to reopening. 

Partisan cracks on reopening are emerging. This general hesitation about reopening, however, masks an emerging partisan divide. Republicans jumped from 9% in support of immediate reopening to 19%, while Democrats barely moved, from 2% in late April to 3% in early May. The net effect of the Republican shift is that a majority of Republicans now prefer reopening in four weeks or less, compared with over four in five Democrats preferring to wait six or more weeks. 

More than 80% of Americans support closing non-essential businesses. Support for limiting restaurants, closing schools, canceling sporting and entertainment events, and group gatherings exceeds 90%. A total of 94% strongly or somewhat approve asking people to stay home and avoid gathering in groups; 92% support canceling major sports and entertainment events; 91% approve closing K-12 schools; 91% approve limiting restaurants to carry-out only; 83% approve closing businesses other than grocery stores and pharmacies. There are some partisan differences on these items -- Republicans are somewhat less supportive, but even among Republicans large majorities support all of these measures; and, as summarized below, support is largely consistent across every state. 

A bipartisan consensus opposes a rapid “reopening” of the economy. Only 7% support immediate reopening of the economy, and the median respondent supports waiting four to six weeks. There is a bipartisan consensus on waiting (89% of Republicans as compared to 96% of Democrats opposed immediate re-opening), and Republicans support a somewhat faster re-opening of the economy than Democrats, where the median Republican supports waiting two to four weeks versus median Democrat six to eight weeks. As discussed below, even in those Republican-led states which are moving toward re-opening, few people support reopening immediately. 

Generally, Americans report adhering to social distancing, indicating that they had minimal social interactions with people outside of their households. That said, 56% reported encountering at least one person from outside of their home in the preceding 24 hours (and 7% reported encountering 10 or more persons); the survey did not contain information on the circumstances of those encounters (e.g., was it at grocery stores? were the individuals wearing masks?). Generally, there were not large differences with respect to age, gender, race, income, partisanship or education. An exception was that Asian Americans were substantially less likely to encounter other individuals, and more likely to avoid contact with other people. There were significant racial differences reported in wearing face masks outside of the home, with 51% of whites reporting following recommendations very closely, along with 62% of Hispanics, 64% of African Americans, and 68% of Asian Americans. There was also an age gradient in this regard, ranging from 50% face mask wearing for 18-24 year olds to 60% of those aged 65 or higher. There were also partisan differences: 51% of Republicans, compared to 64% for Democrats, reported wearing face masks outside the home. 

The most trusted groups and organizations regarding the crisis embody expertise. The Centers for Disease Control (CDC) is the most trusted government entity, with nearly 9 in 10 (88%) indicating that they trusted the CDC “some” or “a lot.” A remarkable 96% indicate trust in hospitals and doctors (only 1% “not at all”); and 93% trusted scientists and researchers. This trust is bipartisan, but there are partisan differences in the degree of trust (i.e., “a lot” versus “some” trust), where 42% of Republicans versus 55% of Democrats report trusting the CDC a lot; 54% of Republicans versus 66% of Democrats report trusting scientists and researchers a lot; and 73% of Republicans versus 76% of Democrats report trusting doctors and hospitals a lot. Less trusted are social media companies (41%) and news media (56%), with Republicans less trusting than Democrats in both cases.