it’s probably ok to send kids to school during COVID: a review of some research about young children and covid

My COVID policy: protect the elderly, release the young. The basis for this recommendation is simple: young people are at very low risk for mortality from COVID. However, a common refrain is “won’t kids catch COVID and take it home.” Well, we now have some evidence that can provide insight.

In Science Magazine, Snape and Viner review a few studies that now focus on the issue of whether small children are responsible for variation in infection rates. Let’s dig into these studies:

First, do kids bring COVID from home to school? Viner’s team has a preprint and the answer is “probably not.” They review the literature to find contact tracing studies that explore how often children were likely to be spreaders vs. adults. The answer? .56 – children are half as likely to be spreaders as adults.

Second, do kids bring COVID from school to home? This is hard to estimate, but Snape and Viner cite recent research on contact tracing and find that school kids are a very small fraction of the people who actually spread COVID. For example, Public Health England study of nasal swabs in British children shows that a tiny fraction of kids at home have active COVID infections. How infrequent? 3.9 out of 100,000.

Of course, this isn’t the last word. But right now, data indicate that children spreaders are not the issue. It’s not zero transmission, which is an incredibly tough standard, but it’s simply small compared to adult transmission. I also note that the studies that Snape and Viner cite use data from the Spring when we were initially reacting to COVID. Today, we know that masks work and keeping physical distance works. Schools are also moving to low density set ups, such as having classes outside when possible and rotating the student attendance (e.g., only 50% of students show up on a given day). Thus, the risk of disease from young people is probably lower in Fall 2020 than Spring 2020.

The conclusion here is simple: opening parts of society aimed at young people is fairly safe. Kids rarely get sick from COVID, they spread less disease than adults, and the prevalence of COVID among kids is really, really low.

50+ chapters of grad skool advice goodness: Grad Skool Rulz ($4.44 – cheap!!!!)
Intro to sociology for just $1 per chapter – INSANE BARGAIN!!!!!
A theory book you can understand!!! Theory for the Working Sociologist (discount code: ROJAS – 30% off!!)
The rise of Black Studies:  From Black Power to Black Studies 
Did Obama tank the antiwar movement? Party in the Street
Read Contexts Magazine– It’s Awesome!!!!

Written by fabiorojas

September 30, 2020 at 12:47 am

Posted in uncategorized

15 Responses

Subscribe to comments with RSS.

  1. Maybe that would be true, if K-5 schools were the only thing open and we didn’t have people dining indoors or college students socializing.



    September 30, 2020 at 4:04 am

  2. Really enjoy your posts in general, Fabio! But I think you have been making some dangerous statements by promoting the “protect the elderly + release the young” policy. I do not claim to be an infectious disease expert. But most experts I have seen emphasize that COVID does not remain confined to a particular age cohort — in the real world any way. Why? Because we can’t simply imprison the elderly. Plus, you seem to be completely ignoring the very significant long-term effects of COVID that many people experience. Here is just one recent link to lack of support for your claim: We are also seeing evidence in Ontario right now that the infections are spreading from younger to older people. So I would encourage you to err on the side of caution.


    Maxim Voronov

    September 30, 2020 at 4:50 pm

  3. The argument is three fold about young people:

    – COVID hospitalization and mortality is very, very rare among the young. In other words, it’s simply low risk. This is different than COVID infection, which happens among all age groups.

    – Contact tracing studies (see the meta study above) don’t often find that the contacts are from school age kids to older adults.

    Also, I did not mention this but Snope and Viner review studies showing that k-12 school opening are not correlated with increases in infection rates.

    Finally, the tweet that Maxim linked to is about infections in two Indian states. My arguments are about symptomatic cases and the evidence is very strong – young people just don’t suffer as much as older people when it comes to COVID.



    September 30, 2020 at 4:58 pm

  4. That is all fine and good: young people don’t suffer very much. But the eveidence that they spread it to the older and more vulnerable people who do suffer much more severe forms of it is simply overwhelming. So how can we release the young, if they will then kill the older people?


    Maxim Voronov

    September 30, 2020 at 6:04 pm

  5. I think you are wrong with your take. There is an emerging consensus that young people can suffer severe damage from COVID-19 infections. We can’t properly estimate the scale of long-term damage to the population right now because we will need to wait a few years. However, the research that is currently being published already shows that this could have huge population level effects and that this can also affect young children and young adults that had seemingly mild disease trajectories.



    September 30, 2020 at 9:12 pm

  6. Anon: I think this argument conflates two issues:

    1. Do young people suffer from COVID?
    2. How often do young people suffer from COVID?

    For me, policy should be about #2. The fact that *some* people get sick is not the whole story. You have to know about frequencies so you can balance the cost of various policies (e.g., if 1,000 kids nation wide die from COVID under policy X, it it worth it? How does that change if it’s 100,000?). Also, many of the links you provide are about the fact that COVID is actually quite damaging to *some* individuals. They do not overturn the general story that *most* COVID illness is disproportionately on the old.

    It is absolutely correct that *some* young people get sick and die. But for policy, it’s about the size of the risk. My argument stands: children rarely show symptoms (but they do get sick – and seriously – in small numbers) but old people have incredibly high rates of serious morbidity and mortality. Furthermore, studies like those cited by Snopes and Viner suggest that child-elderly transmission is rare in contact tracing studies. Thus, we should be focused more on protecting the old.



    September 30, 2020 at 11:15 pm

  7. You only talk deaths – it’s not just about mortality rates but we currently can’t estimate the long-term effects of COVID-19 because the real effects will only be visible in a few years and I don’t see you incorporating this into your analysis. There might also be delayed mortality or morbidity (e.g. some patients have seizures after being asymptomatic, etc.). So yes, of course, fewer young people die but only looking at deaths does not take long-term population health into account (which can be super costly).



    September 30, 2020 at 11:51 pm

  8. You make a fair point, but here is the relevant question. All these long term side effects: are they popping up in asymptomatic people or are they conditional on showing serious symptoms soon after infection?

    If the former is true, then your critique is correct. Lots of people will be COVID and be asymptomatic but have nasty side effects years down the line. And the analysis does not take it into account.

    If the later is true and long term side effects are mainly correlated with clearly presented symptoms, then I am justified in focusing on a proxy of for damaging symptomatic cases, which is mortality, and the conclusion that most costs of COVID are born by the elderly and thus we should focus on them.

    But your other critique about not having time to observe isn’t quite right. The first major COVID wave in the US ran from March to June (see here: At the very least we have a few hundred thousand cases of people who got infected and did not die from Spring 2020. Right about now, we should be getting information. How many people who had COVID 6 months ago and showed no symptoms then have symptoms now?

    If you can show me a ton of young people who had aymptomatic COVID in Spring 2020 but now have severe problems in Fall 2020, I’m with you. But for now, I have seen lots of studies that say “in *some* people, you have long term consequences.” And these people tend to be ones who were sick to start with, and they are concentrated among the elderly. That’s not enough to overturn the general model that (a) lots of people get COVID, but sickness and mortality seem to be strongly correlated with age; (b) distancing and masks really help with reducing transmission and (c) the preliminary evidence that school-child-old person is rare in trace contacting studies. If you add these up, it suggests a more “open” position.

    PS. If I were to really be consistent in counting costs, your point about delayed side effects of COVID supports a more open position. Why? Normally, future costs of a policy are discounted (e.g., death today is worse than death five years from now). Many people might reasonably choose a more “open” society today with more discounted long term effects than a closed situation where you have immediate higher unemployed but fewer long term costs.



    October 1, 2020 at 12:09 am

  9. My main question to you is how, in practical terms, do you intend to release the young people without them infecting the older and more vulnerable population? Unless you create a concentration camp for the elderly and keep the young people away from them, it’s not going to work. We see it in pretty much every jurisdiction that has opened up, and most second waves are driven by young people who are eventually infecting the others. The only other solution is mass testing, and most places in North America are not anywhere close to being able to do this without restricting the freedom of young people.

    In all, nice theory and hard to argue against. But in practical terms, not doable without creating mass casualties.


    Maxim Voronov

    October 1, 2020 at 1:39 am

  10. I’ll write this up in a more detailed post next week, but we have multiple solutions, which range from easy to hard. I claim no originality here as many have been suggested by pro-“open” epidemiologists like Yih and Kulldorff.

    Easy would include age screening at public places, elderly only shopping times, and increased use of delivery services. Medium solutions would be increased testing and immediate quarantine for workers (like nursing home staff) who come into contact with the elderly. Tougher solutions would be to require testing for people who visit nursing homes, testing for people living with elderly people, and increased funding for having home helpers who have tested negative so family members can reduce contact.

    Also, we can informally introduce norms like you should visit elderly people less frequently and elderly people can volunteer to reduce contacts. And elderly people can do things to minimize contact with younger family members. For example, in temperate climates, the elderly could literally just hang out at parks and keep distance rather than stay enclosed at home in close proximity to potentially infected family members. Basically, the elderly would only stay at home to sleep or if they are literally unable to move (in which case, a public policy solution would be to subsidize negative tested home workers – see above).

    From my point of view, preventing the spread of COVID from young people to elderly seems like a challenging but solvable problem with reasonable costs. In contrast, closing all of society, as are finding out, has massive costs – disease does not disappear and we have massive unemployment for adults and school closures for the young. And as most epidemiologists will tell you, unless there is a vaccine, you are just shifting in time at which people will come into contact with the pathogen and develop some degree of immunity.



    October 1, 2020 at 2:57 am

  11. Two points here:

    1. In essence you are suggesting isolating the older and the vulnerable from the rest of the society. I don’t think people want to be isolated from their families and loved ones of a different cohort. Putting the needs of one generational cohort over another is dubious ethics.

    2. More importantly, you are presenting a false choice between open and close. I know that seems to be the discourse in the US, but that is not the totality. The jurisdictions that have been most successful, such as in South Korea and Taiwan have followed a more dynamic model of carefully identifying high transmission risk businesses/locations and shutting them down until the incidence of transmission is lowered and then reopening them. I do not claim to be an expert on infectious diseases. But it seems that the pro-open epidemiologists is a false category, because there are not may pro-close epidemiologists, either. People like David Fisman (UToronto), for instance, advocate for proactive targeted public health measures precisely to avoid the shutdowns of society that you are concerned about. For example, when there are a lot of cases emerging in nightclubs, we can shut them down for a bit, until things are under better control.

    In any case, businesses cannot function well, when there are lots of cases in the community, because the public confidence is necessary for the functioning of the economy. We have seen, for instance, that Sweden did not do much better than its neighbors on the economy front, despite using a much more open approach.

    So in sum:

    – It is nearly impossibly at present to contain COVID within generational cohorts without creating some sort of a morally questionable social stratification between able-bodied and weak-bodied.

    – There is no light-switch for opening or closing the society. Co-existing with COVID for the foreseeable future (even after the vaccine is available) requires governments to listen to public health experts who mostly advocate a dynamic approach of closing and reopening of problem areas in a targeted and strategic manner.


    Maxim Voronov

    October 1, 2020 at 4:35 am

  12. I will delve more into protecting the elderly in a post next week, but I leave you with a few responses this evening:

    1. In general, people who suffer from a disease are the ones who should work the most to prevent getting the disease and they should bear most of the costs of prevention. Even in a world of socialized medicine, if a group X suffers a disease the most, they have the obligation to take reasonable measure to reduce the chance of contracting a disease. If they don’t do it, they are essentially creating more costs for themselves and the rest of society. Asking a group Y that suffers little to bear great cost makes little sense. Thus, I see no problem asking the elderly to take measures to reduce exposure.

    2. Most all the measures I mentioned are actually fairly modest and have relatively low cost on the elderly. Also, they are mostly voluntary. For example, asking the elderly to shop at specific times is simple and very easy to do. Testing family members for COVID frequently is low cost – and makes it easier to maintain contact once you are negative.

    3. Glad you mentioned the epidemiology thing. The standard position in public health pre-COVID was that society wide lock downs and widespread shut downs were very undesirable because of the massive collateral damage. I think this became standard thinking in the 2000s. Also, people didn’t want to delay the rise of herd immunity. So yes, there are very few pro-close epidemiologists – but I had to mention that because a lot of people think that most epidemiologists have always favored wide spread closure. The “textbook” response to epidemics pre-2019 was to combine targeted interventions (quarantines), with lots of contact tracing with various mitigation strategies but allow most social institutions to operate… which is similar to what I propose. The behavior of many actors deviates not only from pre-COVID recommendations, but also from the practice in the 1957 and 1968 flu outbreaks, each of which killed about 100,000 Americans each. To me, the difference in response deserves closer inspection.



    October 1, 2020 at 5:09 am

  13. Please don’t follow this up with future posts as you suggest in the comments. What you’ve said here is already enough of a terrible take. You’re misinterpreting essentially everything you’ve read. You’re selectively ignoring the research that clearly demonstrates why your suggested approach is an awful idea. I’d take the time to rebut what you’ve said point by point, but it’s clear that would be a waste of time. You’ve made it sufficiently clear you have no idea what you’re talking about when it comes to covid. So please, just stop.



    October 1, 2020 at 2:34 pm

  14. @Fabio
    Yes, some of the asymptomatic people develop severe symptoms after the infection (there are some pointers in the articles that I linked previously).

    Also, see

    Quote: “Some of the possible long-term effects can affect even patients who are asymptomatic or have mild cases of COVID-19.

    “I think it’s an argument for why we take this disease so seriously,” says Dr. Poland. “People who are thinking, especially young people: ‘(It’s a) mild disease, you know. I might not even have any symptoms, and I’m over it.’ Whoa. The data is suggesting otherwise. There’s evidence of myocardial damage, cardiomyopathy, arrhythmias, decreased ejection fractions, pulmonary scarring and strokes.”

    Regarding your point of long-term observations: We have a problem of counting here (think about the problem of counting suicides based on death records) because there might be cases where we can’t determine whether the infection caused a stroke or not because it might occur 6 months after the initial infection.
    More data isn’t always helpful and we should think about what is possible, not naively taking current data as “true” or even accurate. The risk is incalculable given the data (this has a nice discussion about different types of “probability” ).

    By the way, Sweden has tried the approach of isolating the elderly as you suggest and it failed miserably. Solutions like this need strong bottom-up support and considering how many people still run around as if nothing happened, I believe they are doomed to fail. With this many seed cases your strategy seems risky.

    I’m not opposed to opening and at least in terms of transmission pathways, we have learned enough to estimate under what conditions disease spread is likely. Considering the incoming data and what we know, schools seem to be a focal point for the spread of COVID-19.



    October 1, 2020 at 4:38 pm

  15. Anon

    October 1, 2020 at 8:13 pm

Comments are closed.

%d bloggers like this: