We know some very important facts about COVID-19, including that vaccines work well. I do not think we understand as much as we should about the effects of government policies on the pandemic. For instance, any medical professional would wear personal protective gear while treating a patient with COVID, but it is less clear that requiring a population to wear masks has any effect. There is often a big slippage between voluntary behavior by trained professionals and large-scale mandates.

I did my own light modeling last April and found no effects of state masking and vaccination mandates on COVID mortality rates. I did find that COVID deaths by US state reflected the percentage of the population who had been in poor health before the pandemic, GOP vote share (Trump support meant higher death rates), Black/White segregation, economic inequality, the percent of the population over age 65, the incarceration rate, and a lower college graduation rate. (Statistically significant correlates: the first three. Adjusted r-square of the whole model = .699).

A new paper by Sun & Biseti examines the effects of state policies on county-level COVID death rates during the first 39 weeks of the pandemic, i.e., before vaccines were available and before masks were being widely recommended. They create a “stringency index” composed of “closures of schools, closures of workplaces, cancellations of public events, restrictions on gatherings, closures of public transport, stay-at-home orders, restrictions on domestic movement, and restrictions on international travel.”

Their model incorporates some similar contextual factors to mine but assesses different policies. It suggests that if every state had employed the maximum of all the stringent measures, the national death rate would have fallen by about 7% in 2020, but the benefits would have been greater “in counties with fewer physicians and larger shares of older adults, low-educated residents, and Trump voters” as well as “in rural areas and counties with higher social capital and larger shares of uninsured residents,” while the benefits would have been smaller “in counties with larger shares of [non-Hispanic] Black and Hispanic residents.” Although I don’t think we can tell from their model itself, it’s plausible that closing schools and businesses had no effect on deaths from the disease in big cities, although the closures were very hard on people.

An older but still valuable article (Sharma, Mindermann & Rogers-Smith 2021) looks at similar measures across subnational units (such as regions or states) of European countries. Their model finds significant benefits from stringent measures such as school closings, but smaller benefits during the pandemic’s second wave than its first. That finding illustrates that we are not in the domain of scientific laws here; we are in a messy zone of rapid change.

In my view, democratic governments and other legitimate institutions have a right to impose many kinds of restrictions to combat a disease. They must do their best to make decisions in the face of uncertainty and conflicting interests. No one’s fundamental human rights have been violated if a government closes schools for months or makes one wear a mask during a pandemic.

On the other hand, this does not mean that the most stringent measures are always effective or that we should be overly confident that we know what works. On the contrary, the lessons of this pandemic appear murky to me, and humility is warranted by all.

A lot of people are very sure what should have been done and are certain their opponents are badly motivated or fools. I think most of us did our best and still don’t have a firm basis to know what we should do next time.

Sources: Sharma, M., Mindermann, S., Rogers-Smith, C. et al. Understanding the effectiveness of government interventions against the resurgence of COVID-19 in Europe. Nat Commun 12, 5820 (2021). https://doi.org/10.1038/s41467-021-26013-4; Sun, Y., & Bisesti, E. M. (2023). Political Economy of the COVID-19 Pandemic: How State Policies Shape County-Level Disparities in COVID-19 Deaths. Socius, 9, 23780231221149902. See also: what explains state variation in COVID-19 mortality?; we must be able to disagree about pandemic policies; vaccination, masking, political polarization, and the authority of science.

In “Who Protests, What Do They Protest, and Why?” (NBER Working Paper 29987), Erica Chenoweth, Barton H. Hamilton, Hedwig Lee, Nicholas W. Papageorge, Stephen P. Roll and Matthew V. Zahn uncover some highly unexpected and challenging findings.

Their data suggest that the people who participated in Black Lives Matter protests in 2020 substantially overlapped with those who protested in favor of reopening schools and businesses during the pandemic. “Attendance at a BLM protest strongly predicts attendance at a Reopening protest.” This finding challenges assumptions about polarization. It also suggests that my local observations in Cambridge, MA were unrepresentative. Here, the most prominent advocates of racial justice were also proponents of closing schools and requiring strong social distancing, but the opposite seems to be closer to the truth across the country.

Chenoweth and colleagues find that “the median protester is white, middle class (measured by income), employed, and a parent.” African Americans are slightly overrepresented in both kinds of protests; but once the authors control for other factors, being Black is correlated with not attending a BLM protest as well as with attending a “reopening” protest. (These associations are small but statistically significant.)

Being “young, low income, [having] young children at home, working in-person, positive beliefs about life, partisanship, higher beliefs of COVID infection, and higher levels of available protests and voter participation predict attendance” at both BLM and reopening protests. Protesters are significantly more likely to vote, which challenges an assumption that protesting and participating in official politics are rival options.

I have quickly explored similar issues using the Tufts Equity in America dataset. It has limitations, and a major one is that we didn’t ask about participation in protests to reopen schools. But we did ask about protest in general, about many opinions regarding COVID-19, and about support for Black Lives Matter–as well as scores of other measures. I used attendance at a protest as the dependent variable in an Ordinary Least Squares regression and chose variables comparable to those in the study by Chenoweth et al. Being female, more educated, and hopeful about the future and knowing someone affected by police violence emerged as positive predictors.

Only a small proportion of our sample was asked about school COVID-19 policies that affected their own children. (They had to be current parents of school-aged children in our 2020 wave). In a very simple regression model, feeling that the school’s policies had been academically detrimental was associated with protesting (where the protests could be on any topic).

See also: Differences in COVID-19 response; Two-thirds of African Americans know someone mistreated by police, and 22% report mistreatment in past year

In “Who Protests, What Do They Protest, and Why?” (NBER Working Paper 29987), Erica Chenoweth, Barton H. Hamilton, Hedwig Lee, Nicholas W. Papageorge, Stephen P. Roll and Matthew V. Zahn uncover some highly unexpected and challenging findings.

Their data suggest that the people who participated in Black Lives Matter protests in 2020 substantially overlapped with those who protested in favor of reopening schools and businesses during the pandemic. “Attendance at a BLM protest strongly predicts attendance at a Reopening protest.” This finding challenges assumptions about polarization. It also suggests that my local observations in Cambridge, MA were unrepresentative. Here, the most prominent advocates of racial justice were also proponents of closing schools and requiring strong social distancing, but the opposite seems to be closer to the truth across the country.

Chenoweth and colleagues find that “the median protester is white, middle class (measured by income), employed, and a parent.” African Americans are slightly overrepresented in both kinds of protests; but once the authors control for other factors, being Black is correlated with not attending a BLM protest as well as with attending a “reopening” protest. (These associations are small but statistically significant.)

Being “young, low income, [having] young children at home, working in-person, positive beliefs about life, partisanship, higher beliefs of COVID infection, and higher levels of available protests and voter participation predict attendance” at both BLM and reopening protests. Protesters are significantly more likely to vote, which challenges an assumption that protesting and participating in official politics are rival options.

I have quickly explored similar issues using the Tufts Equity in America dataset. It has limitations, and a major one is that we didn’t ask about participation in protests to reopen schools. But we did ask about protest in general, about many opinions regarding COVID-19, and about support for Black Lives Matter–as well as scores of other measures. I used attendance at a protest as the dependent variable in an Ordinary Least Squares regression and chose variables comparable to those in the study by Chenoweth et al. Being female, more educated, and hopeful about the future and knowing someone affected by police violence emerged as positive predictors.

Only a small proportion of our sample was asked about school COVID-19 policies that affected their own children. (They had to be current parents of school-aged children in our 2020 wave). In a very simple regression model, feeling that the school’s policies had been academically detrimental was associated with protesting (where the protests could be on any topic).

See also: Differences in COVID-19 response; Two-thirds of African Americans know someone mistreated by police, and 22% report mistreatment in past year

Why have some states seen many more deaths from COVID-19 than others? Do differences in state policies matter? Is it mostly about demographics? Or what about factors like climate and population density, which could influence whether and when people congregate indoors?

To explore these questions, I made a spreadsheet with 58 salient variables about the 50 states, drawing most of the data from the Senate Joint Economic Committee or the Kaiser Family Foundation. I then went fishing for variables that could predict cumulative death rates from COVID-19. I use this “fishing” metaphor with irony, because there is a danger of obtaining spurious results when you explore too many variables at once. Still, the following results might suggest tighter research questions.

Below, I describe nine regression (OLS) models, each with a different thematic focus, arranged in order by how much variance in the states’ COVID-19 mortality they seem to explain. (I report adjusted r-square statistics, which should allow the models to be compared despite differences in the number of variables.)

In summary: the states’ policies that I measured and the partisanship of governors did not matter, but the proportion of people who voted for Trump did. That relationship was not explained by demographics, which I controlled for.

Variables that mattered in many of my models included the percentage of the population that was already in poor health, the GOP vote share in 2020, Black/White residential segregation, and the GINI coefficient (a measure of inequality). A model with just those four components could explain 71% of the variance in COVID deaths (unadjusted r-square = .715).

1. A politics and policy model. Variables: party of state governor, percent of the 2020 state’s popular vote for Republicans, whether the state required masks indoors for some people in Feb 2022, whether the state required, allowed, or banned local vaccine requirements, and state/local spending per capita. The only statistically significant correlate of the mortality rate: the GOP vote share in 2020. Adjusted r-square = .203, meaning that this model offers little insight.
2. A geography model. Variables: population density, percentage rural, average commuting time, mean daily temperature. Statistically significant correlates: none. Adjusted r-square = .240 (again, a poor fit).
3. Sociability model: Variables: average number of close friends, percent of neighbors who regularly do favors, number of nonprofits per 1,000 people, percentage who worked with neighbors to fix/improve something. Statistically significant correlate: working with neighbors (related to lower mortality). Adjusted r-square = .415.
4. A comorbidities model: Variables (all measured pre-pandemic): percent in poor health, premature mortality rate, mortality from suicide/drug overdose, percent disabled, percent with diabetes, obese, and smokers. Statistically significant correlates: general poor health and disabilities. Adjusted r-square = .451.
5. A political participation model: Variables: percent who participated in a demonstration, attended a public meeting, served on a committee, and voted in 2012 and 2016. Statistically significant correlate: attending a public meeting (related to lower mortality). Adjusted r-square = .483.
6. An economics model. Variables: unemployment, incarceration, poverty, GINI coefficient, college graduation rate, internet access at home. Statistically significant correlates: worse inequality, higher incarceration, fewer people with BAs. Adjusted r-square = .623.
7. An inequality model: Variables: Black/White residential segregation, GINI coefficient, college graduation rate, incarceration rate. Statistically significant correlates: racial segregation, GINI coefficient. Adjusted r-square: .646.
8. A politics and demographics model. Variables: the party of state governor, percent of the 2020 state vote for Trump, and the racial demographics and median age of the state. Statistically significant correlates: higher GOP vote, more African Americans, more Latinos, a higher median age. Adjusted r-square = .647.
9. A model that explains most of the variance. Variables: percent in poor health before the pandemic, GOP vote share, Black/White segregation, GINI coefficient, percent over age 65, incarceration rate, college graduation rate. Statistically significant correlates: the first three. Adjusted r-square = .699. (Unadjusted r-square = .735.)

My dataset also included some variables that I have not mentioned here, including several measures of trust (for other people and for institutions) and other types of civic and political participation. None seemed to be influential in any of the models I tried.

The social media and news sources that I follow are full of strong statements about masking rules, vaccine mandates, school closings and other pandemic policies. Some people argue that proponents of loose policies are callous, scientifically ignorant, or even racist because morbidity and mortality rates have been disproportionately high among people of color. Others argue that mandates reflect the arrogance of elites or the creeping power of state bureaucracies. On that side of the argument are some libertarians who would usually be placed on the right, but also some leftist thinkers who are skeptical of science and state power, in the tradition of Horkheimer & Adorno, Michel Foucault, Bruno Latour, Giorgio Agamben, et al. There is also a partisan layer in this debate, with caution about the pandemic being coded as Democratic, and skepticism about its seriousness as Republican.

I rarely depict “both sides” in US politics as equally extreme and polarized. I generally believe that the left wing of the Democratic party represents valid perspectives within a constitutional order while the Trumpian right presents a threat to that order. Still, a recent survey finds, “Nearly half (48%) of Democratic voters think federal and state governments should be able to fine or imprison individuals who publicly question the efficacy of the existing COVID-19 vaccines on social media, television, radio, or in online or digital publications.” This statistic comes from a right-leaning pollster. I don’t have any reason to doubt the concrete result, but I would have investigated possible intolerance on the other side of the debate as well. I would guess that significant numbers of respondents would support locking up school boards that mandate masks and prosecuting Dr. Fauci. Meanwhile, some serious writers on both sides reject the legitimacy of disagreement and use opposing arguments about COVID-19 as evidence that our whole political system is fundamentally broken.

Our system may indeed be close to breaking down, but not because individuals have the temerity to disagree about COVID-19 policies.

A caveat: it is not clear that the real debate is as hot, personalized, and divided as my media feed suggests. Twitter attracts controversialists with strong, ideological perspectives, whereas many Americans are apolitical. The news media covers controversy and gives little attention to routine decision-making. Outrageous threats at a school board meeting can attract national attention while a boring agreement will draw low-key local coverage, at most. However, there are plenty of serious people who publicly deny the legitimacy of disagreement about COVID-19, and they require a response.

I would start with a general view of politics. All types and layers of governments and other institutions–including firms–constantly make grave decisions. They imprison people, fire them, and give or deny them crucial services. Even routine decisions, such as zoning regulations or the development of new products, can profoundly affect people’s welfare. Although some decisions are simply good or bad, many are debatable. They have both winners and losers, they involve conflicting values, and their consequences are unpredictable. Nor is it safe to do nothing, for that can sometimes be a harmful failure.

Americans don’t particularly like disagreement, especially when it involves conflicts of principle and identity under conditions of uncertainty. Therefore, we place many consequential decisions out of view. For instance, we have dramatically reduced the number of jury trials (which require regular citizens to make choices) in favor of plea-bargaining. And decisions about matters like zoning are made in forums that draw very little attention.

COVID-19 has forced such decisions into the open. Like other issues, it involves conflicting values and interests under uncertain conditions. Yes, vaccines are highly effective and safe, and critics do themselves no credit when they doubt such findings. A large, randomized, double-blind experiment with a mass-produced chemical product presents an exceptional opportunity to resolve empirical uncertainty. However, there is plenty of room for doubt about the empirics of other matters, such as school closings and masks, and even about mandates for vaccines.

Indeed, the evidence about the effects of policies on the pandemic is murky. You can tell it’s confusing just by glancing at the ten states with the highest per capita cumulative death rates so far, which include Mississippi and Oklahoma but also New Jersey and New York. (Among the best-off so far: Utah and Nebraska as well as Vermont and Hawaii.) Of course, one should control for factors other than state policies. A typical study that uses controls finds small effects: e.g., mask mandates reduce the growth of cases by 2 percentage points. I think that finding counts in favor of a mask mandate, but with many caveats; it certainly does not neutralize all concerns or close the case. For one thing, the virus itself keeps changing, as do other circumstances, such as the percentage of people with immunity. Also, the pandemic has rolled out in regional waves, which means that the same methodology will yield different results depending on when the study is conducted. We won’t have a clear picture until it is clearly over.

If you believe in democracy, you should be glad that people can influence public decisions. If you believe in pluralism or polycentricity, you should be glad that there are many different forums for decision-making: federal, state, and local governments; executive, legislative, and judicial bodies; corporations and nonprofit institutions; professional and scientific bodies; and transnational organizations. You should see disagreement as evidence of liberty, diversity, and participation.

But you won’t get the policies you want. If you’re fortunate, you may be aligned with public opinion and the decision-makers in your own community. Then you will appreciate local policies and will probably observe reasonably high levels of voluntary compliance. However, in a polycentric world, you will not see the policies you support enacted or obeyed everywhere else. Communities will vary. Yet the policies adopted in other places may affect you. So the variation will be frustrating and even angering.

People are entitled to strong views and emotions, including anger. But it is important to distinguish process from outcomes. State and local governments in the US may decide whether to require masks or not. Some decisions may be wiser than others, but the unwise ones are still legitimate. If some people have to wear mask when they don’t believe in them–or attend schools where masks are absent even though they do believe in them–that is democracy at work. The health risks may be serious, but governments constantly make decisions that affect health, and even life. People walking around in mandatory masks are not serfs to a tyrannical state, but communities that have eschewed masks are not idiotic. We disagree. Decisions must be made. It is good that we the people can make them.

Here are some tips to consider:

• Don’t threaten or bully individuals. Certainly, do not try to jail them for their opinions.
• Obey politically legitimate policies even if you disagree with them unless they violate your core principles, but be careful about mistaking your opinions for sacred principles. Usually, decisions require some to compromise what we want and believe.
• If you are on the winning side, acknowledge that the losing side is being asked to sacrifice.
• Protect others’ freedom of speech, not only from censorship but also from the tyranny of majority opinion.
• Pay attention to equity and structural forms of injustice, but don’t assume that you know what people believe (or what is good for them) based on their demographics.
• To address scientific issues, look for the most recent and rigorous scientific publications. Googling around for opinions is not “research.” On the other hand, do not overstate the policy significance of specific scientific papers, and do not use empirical findings to squelch normative disagreements. For instance, if mask mandates reduce the spread of COVID-19, it does not automatically follow that a state’s governor should require masks in all public schools.
• Hold onto your general political and philosophical views (if you wish), but don’t use the pandemic as an opportunity to score debating points on behalf of your philosophy. We should be trying to do the right thing here and now. Besides, the current pandemic may prove more of an exception than a proof-point for several leading ideologies. Libertarians should recognize that libertarian thinkers have often endorsed restrictions during epidemics. Critics of mainstream science should acknowledge the enormous value of the corporate-produced vaccines. Progressives (like me) should ask why well-funded public scientific agencies have performed so poorly in several respects.
• Keep an eye open for arguments and evidence that trouble your own assumptions, but don’t give up on trying to decide what’s really best to do under the circumstances, with the evidence that we have at hand.

See also: collected posts on the COVID-19 pandemic, and in particular, vaccination, masking, political polarization, and the authority of science; why protect civil liberties in a pandemic?; and theorizing democracy in a pandemic.