COVID Lockdowns Were a Giant Experiment. It Was a Failure. – New York Magazine

June 10, 2020, in Williamsburg. Photo: JOHN TAGGART/The New York Times/REDUX

On April 8, 2020, the Chinese government lifted its lockdown of Wuhan. It had lasted 76 days two and a half months during which no one was allowed to leave this industrial city of 11 million people, or even leave their homes. Until the Chinese government deployed this tactic, a strict batten-down-the-hatches approach had never been used before to combat a pandemic. Yes, for centuries infected people had been quarantined in their homes, where they would either recover or die. But that was very different from locking down an entire city; the World Health Organization called it unprecedented in public health history.

The word the citizens of Wuhan used to describe their situation was fengcheng sealed city. But the English-language media was soon using the word lockdown instead and reacting with horror. That the Chinese government can lock millions of people into cities with almost no advance notice should not be considered anything other than terrifying, a China human rights expert told The Guardian. Lawrence O. Gostin, a professor of global health law at Georgetown University, told the Washington Post that these kinds of lockdowns are very rare and never effective.

The Chinese government, however, was committed to this zero-COVID strategy, as it was called. In mid-March 2020, by which time some 50 million people had been forced into lockdowns, China recorded its first day since January with no domestic transmissions which it offered as proof that its approach was working. For their part, Chinese citizens viewed being confined to their homes as their patriotic duty.

For the next two years, harsh lockdowns remained Chinas default response whenever there was an outbreak anywhere in the country. But by March 2022, when the government decided to lock down much of Shanghai after a rise in cases in that city, there was no more talk of patriotism. People reacted with fury, screaming from their balconies, writing bitter denunciations on social media, and, in some cases, committing suicide. When a fire broke out in an apartment building, residents died because the police had locked their doors from the outside. And when the Chinese government finally abandoned lockdowns an implicit admission that they had not been successful in eliminating the pandemic there was a wave of COVID-19 cases as bad as anywhere in the world. (To be fair, this was partly because China did such a poor job of vaccinating its citizens.)

One of the great mysteries of the pandemic is why so many countries followed Chinas example. In the U.S. and the U.K. especially, lockdowns went from being regarded as something that only an authoritarian government would attempt to an example of following the science. But there was never any science behind lockdowns not a single study had ever been undertaken to measure their efficacy in stopping a pandemic. When you got right down to it, lockdowns were little more than a giant experiment.

June 16, 2020, in Bloomington, Minnesota. Photo: Kerem Yucel/AFP via Getty Images

March 25, 2020, in Chicago. Photo: Taylor Glascock/The New York Times/REDUX

Despite the lack of scientific evidence, lockdowns didnt come out of nowhere, at least not in the U.S. They had been discussed and argued over by scientists since 2005, when (as the story goes) President George W. Bush read John M. Barrys book The Great Influenza, about the 1918 pandemic. This happens every hundred years, Bush is supposed to have said after finishing the book. We need a national strategy.

In fact, there were people thinking about pandemic mitigation long before Bush read Barrys book. The leader of this ad hoc group was D.A. Henderson, perhaps the most renowned epidemiologist of the 20th century the man who, decades earlier, had led the team that eradicated smallpox. Richard Preston, the author of The Hot Zone, would later describe this feat as arguably the greatest life-saving achievement in the history of medicine.

By the time Bush began pushing his administration to come up with a pandemic plan, Henderson was 78 years old. Ten years earlier, he had sat in on a series of top-secret briefings where he listened to a Russian defector describe how he had led a team that was trying to adapt the smallpox virus for bioweapons. Henderson became so concerned that he started a small center focused on biodefense which meant, in effect, defending against a pandemic. He and his colleagues at the center had spent years trying to persuade government officials to take pandemics seriously without much success. When the Bush administration began debating what its pandemic strategy should include, it was only natural that Henderson be involved.

The men Bush chose to lead the effort believed that lockdowns could be an important component of a mitigation plan. They were heavily influenced by a model developed by Laura Glass, a 14-year-old high-school student from Albuquerque (aided by her scientist father), that purported to show that keeping people away from one another was as effective as a vaccine. (This story is told, overexcitedly, in Michael Lewiss book The Premonition.)

Henderson vehemently disagreed. For one thing, he didnt trust computer models, which churned out estimates based on hypotheticals. Just as important, they couldnt possibly anticipate the complexity of human behavior. There is simply too little experience to predict how a 21st-century population would respond, for example, to the closure of all schools for periods of many weeks to months, or the cancellation of all gatherings of more than 1,000 people, he said.

In addition, he felt that the worst thing officials could do was overreact, which could create a panic. In 2006, as the debate inside the Bush administration was nearing its conclusion, he co-authored a paper in a final effort to change the minds of those devising the strategy. The paper concluded: Experience has shown that communities faced with epidemics or other adverse events respond best and with the least anxiety when the normal social functioning of the community is least disrupted. Strong political and public health leadership to provide reassurance and to ensure that needed medical care services are provided are critical elements. If either is seen as less than optimal, a manageable epidemic could move towards catastrophe.

The Bush teams final document, published by the CDC in February 2007, stopped short of mandating lockdowns but came as close as its authors dared, calling for the use of social distancing measures to reduce contact between adults in the community and workplace. One of the leaders of the effort, a government scientist named Richard Hatchett, would later tell Lewis what he really believed: One thing thats inarguably true is that if you got everyone and locked each of them in their own room and didnt let them talk to anyone, you would not have any disease.

Which is true as far as it goes. There were other questions, though, that were at least as relevant. Could the kinds of lockdowns that are achievable in the real world, where hundreds of millions of people cant live in isolation chambers, be an effective tool against a pandemic? Did the virus truly go away during a lockdown or simply hide, waiting to reemerge when it ended? And finally, did the many social, economic, and medical downsides make them, in the aggregate, not worth whatever short-term benefits they might yield?

Henderson, who died in 2016, never stopped making the latter case. D.A. kept saying, Look, you have to be practical about this, says his former deputy, Tara OToole. And you have to be humble about what public health can actually do, especially over sustained periods. Society is complicated, and you dont get to control it.

As the United States gains more and more distance from the COVID pandemic, the perspective on what worked, and what did not, becomes not only more clear, but more stark. Operation Warp Speed stands out as a remarkable policy success. And once the vaccines became available, most states did a good job of quickly getting them to the most vulnerable, especially elderly nursing-home residents.

Unfortunately, there is no shortage of policy failures of which to take stock. We do an accounting of many of them in our new book, The Big Fail. But one that looms as large as any, and remains in need of a full reckoning in the public conversation, is the decision to embrace lockdowns. While it is reasonable to think of that policy (in all its many forms, across different sectors of society and the 50 states) as an on-the-fly experiment, doing so demands that we come to a conclusion about the results. For all kinds of reasons, including the countrys deep political divisions, the complexity of the problem, and COVIDs dire human toll, that has been slow to happen. But its time to be clear about the fact that lockdowns for any purpose other than keeping hospitals from being overrun in the short-term were a mistake that should not be repeated. While this is not a definitive accounting of how the damage from lockdowns outweighed the benefits, it is at least an attempt to nudge that conversation forward as the U.S. hopefully begins to recenter public-health best practices on something closer to the vision put forward by Henderson.

February 3, 2020, in Wuhan, China. Photo: Stringer/Getty Images

After China came Italy, the second country to be hit hard by the coronavirus. The Italian government responded with a lockdown almost as tough as Chinas. By the time it was lifted, in early June, 34,000 Italians had died of COVID-19, up from 630 when the lockdown was first imposed.

Those were frightening numbers. But when Neil Ferguson saw what had transpired in Italy, he saw an opportunity. For Ferguson, the head of the infectious disease department at Imperial College London, the Italian governments decision to follow Chinas example meant that lockdowns were suddenly a real-world policy option in Western democratic societies, not just in an authoritarian country like China. As a disease modeler, he believed the same thing Richard Hatchett believed: that if he could lock everyone in a room, the virus would go away. But he had long assumed attempting to do so was politically impossible.

Ferguson is an important epidemiologist, renowned for his estimates, derived from computer models, of possible deaths from a newly emerged virus.As soon as he learned of the outbreak in Wuhan, he and several colleagues began modeling the coronavirus. On March 17, Ferguson laid out the teams findings at a press conference. Their model predicted that, without serious countermeasures, a staggering 81 percent of the population in the U.S. and Britain would become infected, and that 510,000 people in Britain and 2.2 million Americans would die of COVID by late 2020. In addition, the authors wrote, We predict critical care bed capacity would be exceeded as early as the second week in April, with an eventual peak in ICU or critical bed care demand that is over 30 times greater than the maximum supply in both countries.

For Ferguson, the purpose of the report wasnt just to release their shocking estimates; it was also to push the American and British governments to commit to lockdowns for the long haul. [T]his type of intensive intervention package, the authors wrote, will need to be maintained until a vaccine becomes available (potentially 18 months or more) given that we predict that transmission will quickly rebound if interventions are relaxed. It worked. In the U.K., Prime Minister Boris Johnson had initially planned to keep the country open. Instead, he ordered a lockdown within a week of Fergusons press conference. (Shortly after Johnson imposed the lockdown, Ferguson was visited twice by his mistress. For obvious reasons, this caused a furor when it was discovered. Ferguson was the first, though hardly the last, Establishment bigwig to ignore the COVID-19 rules they demanded of everyone else.)

As for President Donald Trump, he never used the word lockdown, but he was worried enough to call for the country to adopt social distancing as a mitigation strategy. Schools, restaurants, businesses they all closed. White-collar employees who were able to work from home did so. More than once, Trump mentioned that 2.2 million lives were at stake, referring to Fergusons estimate. Trumps order wound up lasting six weeks.

Most governors issued their own stay-at-home orders, usually stricter than Trumps. Even Governor Ron DeSantis in Florida who would soon become an outspoken opponent of mainstream mitigation measures reluctantly went along for a brief period. But there were important questions that no one advocating for lockdowns addressed, maybe because in the urgency of the moment the questions didnt occur to them. How long would they last? And even if lockdowns did slow the viruss progression, what would happen when they were lifted?

Regardless, in the space of two months, lockdowns had gone from being unthinkable to being an unquestioned tool in the pandemic toolkit.

March 17, 2020, in Paris. Photo: Veronique de Viguerie/Getty Images

July 5, 2020, in Melbourne. Photo: Asanka Ratnayake/Getty Images

When state public health officials explained to the countrys governors why lockdowns were necessary, they talked primarily about bending or flattening the curve. And when governors then explained the strategy to their constituents, they used the same rationale. If we change our behaviors, said California governor Gavin Newsom in announcing his states lockdown on March 19, we can truly bend the curve to reduce the need to surge. The day after Newsom, then-Governor Andrew Cuomo announced a lockdown plan for New York. He called it his PAUSE program Policies Assure Uniform Safety for Everyone but really, it was the same thing.

What did flattening the curve mean? Heres what it didnt mean: It did not mean that if people stayed in their homes, COVID-19 would fade away (even if that idea was often suggested in non-expert contexts). Rather, flattening the curve meant delaying the virus spread to prevent hospitals from becoming overwhelmed with COVID patients. During their early press conferences, many governors would display a chart showing a sharp increase in the estimated rate of COVID-19 infections. Thats what would happen without lockdowns, they explained. Then they would display a second chart showing a more gradual upturn once lockdowns and other mitigation measures took effect. Simply put, flattening the curve was about helping hospitals manage the crisis rather than ending the crisis. Even those who later criticized lockdowns largely agreed on this point. As David Nabarro, the World Health Organizations COVID-19 envoy (and an eventual lockdown critic), put it, The only time we believe a lockdown is justified is to buy you time to reorganize, regroup, rebalance your resources, protect your health workers who are exhausted, but by and large wed rather not do it.

In many blue states, however, that rationale was forgotten over time, and many people remained confined to their homes or apartments not just for a few weeks but for a year or more even after the vaccine became available. And many of the countrys biggest cities continually reimposed lockdowns whenever there was an uptick in COVID cases not just telling people to shelter in place, but also closing small businesses and restaurants, outlawing sports events and social gatherings, and shutting down in-school learning.

Which naturally leads to the obvious question: Did lockdowns help keep Americans alive? Studies were mixed in their findings, their methodology, even their definition of lockdown. For instance, in August 2020, eClinicalMedicine, an offshoot of the prestigious British medical journal The Lancet, printed a study that concluded that full lockdowns and wide-spread COVID-19 testing were not associated with reductions in the number of critical cases or overall mortality. In March 2021, Christian Bjrnskov, an economist at Aarhus University in Denmark, compared weekly mortality rates in 24 European countries that used mitigation measures with varying degrees of severity. [T]he findings in this paper suggest that more severe lockdown policies have not been associated with lower mortality, the economist wrote. In other words, he added, the lockdowns have not worked as intended.

Michael Osterholm, the prominent epidemiologist at the University of Minnesota, also doesnt think lockdowns did any good. There is actually no role for lockdowns, he says. Look at what happened in China. They locked down for years, and when they finally relaxed that effort, they had a million deaths in two weeks. As for flattening the curve, thats not a real lockdown, Osterholm says. Youre just reducing contact for a few weeks to help the hospitals.

Dr. Anthony Fauci was probably the best-known defender of lockdowns as a life-saving measure. But the policy continues to have many defenders within the public health establishment. Howard Markel, a doctor and medical historian at the University of Michigan, believes they succeeded. The amount of lives saved was just incredible, he says. Markel pointed to an August 2023 study by the Royal Society of London that concluded that stay-at-home orders, physical distancing, and restrictions on gathering size were repeatedly found to be associated with significant reduction in SARS-CoV-2 transmission, with more stringent measures having greater effects.

Still, the weight of the evidence seems to be with those who say that lockdowns did not save many lives. By our count, there are at least 50 studies that come to the same conclusion. After The Big Fail went to press, The Lancet published a study comparing the COVID infection rate and death rate in the 50 states. It concluded that SARS-CoV-2 infections and COVID-19 deaths disproportionately clustered in U.S. states with lower mean years of education, higher poverty rates, limited access to quality health care, and less interpersonal trust the trust that people report having in one another. These sociological factors appear to have made a bigger difference than lockdowns (which were associated with a statistically significant and meaningfully large reduction in the cumulative infection rate, but not the cumulative death rate).

In all of this discussion, however, there is a crucial fact that tends to be forgotten: COVID wasnt the only thing people died from in 2020 and 2021. Cancer victims went undiagnosed because doctors were spending all their time on COVID patients. Critical surgeries were put on hold. There was a dramatic rise in deaths due to alcohol and drug abuse. According to the CDC, one in five high-school students had suicidal thoughts during the pandemic. Domestic violence rose. One New York emergency-room doctor recalls that after the steady stream of COVID patients during March and April of 2020, our ER was basically empty. He added, Nobody was coming in because they were afraid of getting COVID or they believed we were only handling COVID patients.

So in attempting to gauge the value of lockdowns, the most appropriate way is to look not just at COVID deaths but at all deaths during the pandemic years. Thats known as the excess deaths a measure of how many more people died than in a normal year. One authoritative accounting was compiled by The Spectator using data gathered by the OECD. It showed that during the first two years of the pandemic 2020 and 2021 the U.S. had 19 percent more deaths than it normally saw in two years time. For the U.K., there was a 10 percent rise. And for Sweden one of the few countries that had refused to lock down its society it was just 4 percent. An analysis by Bloomberg found broadly similar results. In other words, for all the criticism Sweden shouldered from the worlds public health officials for refusing to institute lockdowns, it wound up seeing a lower overall death rate during the pandemic than most peer nations that shut down schools and public gatherings. It is not unreasonable to conclude from the available data that the lockdowns led to more overall deaths in the U.S. than a policy that resembled Swedens would have.

March 20, 2020, in Los Angeles. Photo: AGUSTIN PAULLIER/AFP via Getty Images

There were other negative consequences too. In the U.S., lockdowns forced hundreds of thousands of small business closures. They exacerbated inequality, as Amazon warehouse workers and meatpackers showed up to crowded workplaces while the Zoom class locked down at home. Worst of all, though, it had a devastating effect on children whose schools were closed as part of a lockdown. During the first weeks of the pandemic it probably made sense to close schools given how little was known about the coronavirus. Better safe than sorry. But by the time school started up again in the fall of 2020, two things were clear. The first was that remote learning was a disaster. The second was that there was surprisingly little transmission among kids in school. Well-to-do parents moved their children to private schools, many of which reopened their classrooms. But most big-city public-school systems continued to rely on remote learning well into the 20202021 school year. It was a tragic policy choice.

In ProPublica and The New Yorker, the journalist Alec MacGillis vividly described the consequences in Baltimore. With no classrooms to go to, thousands of students abandoned school. The school system made free laptops available, but few students took the trouble to get one. Teachers gave up trying to prod those who didnt log onto their remote classes. Plus, teachers had kids of their own to take care of, which made it difficult to teach.

The anti-lockdown scientist Jay Bhattacharya of Stanford University recalls a photograph in the San Jose Mercury News during the early months of the pandemic. It showed two children, 7 or 8 years old, sitting with Google Chromebooks outside a Taco Bell. They were on the sidewalk doing schoolwork because that was the only place they could get free Wi-Fi, Bhattacharya said. Their parents werent there because they had to go to work. I mean, that should have ended the lockdown right then and there. It should have at least ended school closures.

Public schools have an importance that goes beyond education. Its where many of the rituals of childhood and young adulthood take place. For children who live in unstable homes, school offers some stability. Public schools serve free breakfast and lunch to disadvantaged kids. And theyre a place where parents know their children are safe when theyre at work. One consequence of lockdowns was that millions of children had to fend for themselves because their parents couldnt afford to quit their jobs to take care of them.

One child psychiatrist, who works with underprivileged autistic kids, began the pandemic believing in the importance of lockdowns and other mitigation measures. But over time, she changed her mind.

What really drove me was my clinical experience, she said. What happens to a child when every single support is removed from them? Whats the impact on the family and the siblings? What I was seeing was complete regression. It was devastating, and the downsides of lockdowns and school closings were not being openly discussed in the mainstream media. I was horrified. Why arent we talking about this? She described the situation she saw as 2022 wore on as a sickening mental-health crisis.

The science also weighed heavily in favor of opening schools. By mid-summer 2020, when cities were trying to decide whether to reopen schools in September, 146,000 Americans had died of COVID-19. Fewer than 20 were children between the ages of 5 and 14. More schoolchildren died from mass shootings in a typical year. Emily Oster, a Brown University economist, conducted a survey of about 200,000 children who were back in classrooms. The infection rate, she discovered, was 0.13 percent among students and 0.24 percent among teachers an astonishing low number. Oster then set up what she called the National COVID-19 School Response Dashboard, which eventually tracked 12 million kids in both public and private schools and continued to collect infection-rate data over the next nine months. Not once did the student rate hit one percent during any two-week span.

We do not want to be cavalier or put people at risk, Oster wrote in The Atlantic. But by not opening, we are putting people at risk, too.

March 15, 2020, in Manhattan. Photo: VICTOR J. BLUE/The New York Times/REDUX

Over the entirety of the pandemic, the essential facts about schools never changed. The infection rate for teachers in Sweden, where most schools stayed open, was no higher than the infection rate for teachers in Finland, which had closed its schools. In early 2021, three CDC scientists acknowledged in the Journal of the American Medical Association: As many schools have reopened for in-person instruction in some parts of the U.S. there has been little evidence that schools have contributed meaningfully to increased community transmission.

So why did so many big-city schools stay closed long after the evidence was clear? There were three reasons. The first, and most understandable, was fear. No matter how small the chance, no parent wanted his or her child to die from COVID-19. And no teacher wanted to become infected while in school and bring COVID-19 home. Because kids often brought colds and flus to school which then spread to others both parents and teachers had a hard time accepting that that was not how the virus spread. Here, for instance, was a typical comment from a teacher in Westchester, reacting to a series of New York Times articles about reopening schools:

Tell me how to get a 6-year-old to not sneeze on his friends let alone play and work from a distance (mucus, saliva, pee, poop, this is all part of our day at the lower levels of education). Tell me how each child is going to have her own supplies for the day as shared supplies are no longer an option. No more Legos, no more books. Tell me how to comfort a hysterical child from a distance of six feet.

That it was well established that the coronavirus was not spread through saliva or pee or by sharing books didnt matter. Too many people were simply unable to judge risk rationally a problem due in part to unwillingness of government officials to talk honestly about COVID-19. In 2020, for instance, COVID-19 ranked below suicide, cancer, accidents, homicide, and even heart disease as a cause of death for children under the age of 15, according to CDC data. Yet public-health experts did not stress any of this on the contrary, many of them emphasized instead that children could get COVID-19 without explaining how small the risk was. Is it any wonder, then, that COVID-19 seemed to be the only thing parents and teachers focused on?

The second factor was Trump. On July 6, he tweeted, SCHOOLS MUST REOPEN IN THE FALL!! The next day, at a White House event, the president said, Were very much going to put pressure on governors and everybody else to open the schools. Its very important for our country. Its very important for the well-being of the student and the parents.

In this case, Trump happened to be right; it was important. But by this late stage in his presidency, most Democrats assumed that anything he said was a lie. If Trump said schools should reopen, that was reason enough for them to assume they should stay closed. The sense that opening up was a Trump-endorsed policy seems to have energized opposition to it in blue America even as data accumulated that the harm being done to the countrys children outweighed any potential benefit.

The third reason was the teachers unions. Public-school teachers were unionized, and their unions American Federation of Teachers (AFT) and the National Education Association (NEA) were allies of, and contributors to, the Democratic Party, which dominated most major urban areas. They held enormous sway over big-city school systems.

No one can doubt that teachers were afraid of dying of COVID-19. They truly believed they were putting themselves in harms way if they went back into a classroom full of children. But instead of helping their members see how small the risk truly was, the teachers unions embraced the fight to keep teachers away from the classroom.

By the time September 2020 rolled around, at least a dozen of Americas biggest cities started the school year remotely. They included Los Angeles, San Francisco, New York, Chicago, and Houston all cities with the kind of large, disadvantaged communities that would suffer the most if schools were closed. In most cases, city officials said they were trying to move from remote learning to at least a hybrid model, in which students would spend several days a week in classrooms and the rest of the week online.

In school districts that did open their schools that fall, the results were remarkably aligned with Emily Osters data. In New York, Mayor Bill de Blasio was finally able to get the schools open in late September; between Thanksgiving and the end of the year, the citys positivity rate rose from 3 percent to 6 percent. The positivity rate in the public schools also rose from 0.28 percent to 0.67 percent. The safest place in New York City is, of course, our public schools, said de Blasio. To the holdout unions, those numbers didnt matter. Ultimately, only 15 percent of school districts offered full-time classroom instruction during the fall 2020 semester.

By 2022, journalists, academics, and even some public-health officials were finally coming to grips with the enormous damage done to children especially disadvantaged children because of remote learning. A lengthy analysis by two professors in The Atlantic toted up some of the issues. First, millions of kids simply gave up on learning. In New York, even after schools had reopened, the chronic absentee rate was 40 percent up from 26 percent before the pandemic. Studies showed that public-school children got less exercise (no recess) and ate more junk food (no free hot meals) during the pandemic. According to a CDC survey, during the first six months of 2021, nearly half the high-school students surveyed felt persistently sad or helpless. Parental emotional abuse was four times higher than in 2013, and parental physical abuse nearly doubled, The Atlantic reported.

A study by three major research institutions, including Harvards Center for Education Policy Research, showed that the longer a school relied on remote learning, the further behind their students were. In high-poverty schools that were remote for more than half of 2021, the loss was about half of a school years worth of typical achievement growth, said Thomas Kane, the director of the Harvard center.

Although test scores in 2023 would suggest that students were slowly catching up, those scores didnt take into account the kids who had dropped out entirely. One analysis found an estimated 230,000 students in 21 states whose absences could not be accounted for. They had simply gone missing. The pandemic has amounted to a comprehensive assault on the American public school, concluded the authors of The Atlantic article. Yet as late as the fall of 2022, there were still those who refused to acknowledge the damage done by lengthy school closings.

One such person was Anthony Fauci. In August 2022, Fauci announced that he planned to retire at the end of the year. Over the next few months, he made the rounds to discuss how the country had fared during the pandemic. Invariably, he was asked whether he regretted his forceful advocacy of lockdowns, especially given its effect on children. At one forum, he said, Sometimes when you do draconian things, it has collateral negative consequences on the economy,on the schoolchildren. But, he added, The only way to stop something cold in its tracks is to try and shut things down.

What he could never acknowledge was that shutting things down didnt stop the virus, and that keeping schools closed didnt save kids lives. Then again, to understand that, you had to be willing to follow the science.

Excerpted fromThe Big Fail: What the Pandemic Revealed About Who America Protects and Who It Leaves Behind, by Joe Nocera and Bethany McLean (Portfolio, October 2023).

Go here to see the original:

COVID Lockdowns Were a Giant Experiment. It Was a Failure. - New York Magazine