Ricochet is the best place on the internet to discuss the issues of the day, either through commenting on posts or writing your own for our active and dynamic community in a fully moderated environment. In addition, the Ricochet Audio Network offers over 50 original podcasts with new episodes released every day.
It Doesn’t Matter if Masks Work
I’ve been trying to stay out of this. I really have. But just can’t help myself. I’ve read every possible opinion about preventing the spread of COVID using masks. Some people say they don’t work at all. Others say that they work extremely well. Others say they work, but they don’t work because you morons can’t use them correctly. And there are many other variations out there. I haven’t posted on this topic, partially because I don’t know what I’m talking about. I don’t really understand what just happened. So much of this makes no sense. But I can no longer restrain myself. Allow me to propose my own theory: Nearly all the theories I’ve read are wrong. All of them. Well, pretty much. In my view, at least.
The use of statistics in medicine is complex. There are lots and lots of variables, most of which are unknowable. Discerning the impact of a treatment from random chance can be tricky. We try to follow a wide range of studies, but they often don’t agree. But let me try to illustrate my point at first by moving away from masks, and talk briefly of heart disease. This will seem off-topic at first, but you’ll see where I’m going soon. And the doctors in the audience may catch me over-simplifying the data a bit. I’m just trying to explain a general concept here, so if you send me a personal message that involves p-values, I promise I won’t read it. And this post is covering some complex topics, so it will be longer than my usual post. If you can’t stand the idea of reading anything else about masks, I won’t blame you if you just keep scrolling.
Ok. Because heart attacks are often fatal, we spend a lot of time and effort trying to prevent them, rather than treating them after the fact. There are two types of prevention, in the cardiovascular world. There is primary prevention, in which we take people who have not had a heart attack, and we give them some form of treatment, in an effort to avoid their first heart attack. Then, there is secondary prevention, in which we take people who have had a heart attack, and we give them some sort of treatment, in an attempt to prevent subsequent heart attacks.
Let’s start with secondary prevention. If you take 1,000 people who have had had a heart attack, and you give them all statin drugs, will you prevent subsequent heart attacks? Yes. The data is simply overwhelming. Statins work extremely well in this environment, all the studies show pretty much the same thing, and there is no longer any debate on their effectiveness in treating heart patients. In fact, it has become unethical to do placebo trials on such patients, because it is so obvious that heart patients live longer when you give them statins, you just can’t deny half the people in the study a treatment that you know will prolong their life. You are not allowed to intentionally kill people while doing research. If you know a drug works you can’t deny it to half the people in the study. So, yes, statins work extremely well when given to people who have already had a heart attack.
Ok. What about primary prevention? Well, in primary prevention studies, the data on statins is much more complex. Some studies show reasonable benefit, others show small benefits, and some show little benefit at all. While secondary prevention studies look very similar to one another, primary prevention studies are all over the map.
This seems odd, doesn’t it? You’re using the same drug (statins) to treat the same disease (heart attacks), but you’re getting different results in primary prevention vs secondary prevention studies. They should show the same thing, right?
Can you figure out why there are such striking differences between the two types of studies? I’ve given hundreds of speeches to various doctors’ groups across the country on this topic, and I would be embarrassed to tell you how many physicians in my audiences can’t figure out the difference. Do you know what it is?
Since the treatment is the same, and the disease is the same, the difference must be the patients, right?
Think of it like this. In a secondary prevention study, all of the patients have heart disease, and you know it. They have had a heart attack. You don’t need a cholesterol panel or a blood pressure check to figure out if they are at risk for heart disease. You know they have heart disease.
Now, suppose that statins reduced risk of heart attack by 50%, just for simple math. If all of the patients in the treatment group have active heart disease, you might hope that your prevention numbers might get as high as 50%, right? It won’t be that high, of course, because not everyone who has a heart attack has a second heart attack. But many do. So your reduction of risk and heart attacks looks pretty good in that population because they all have heart disease.
Ok, so now think of a primary prevention study. You’re going to take 1,000 people who have high cholesterol, and you’re going to treat them all with a statin, and you’re going to see how many heart attacks you can prevent. Suppose you’re using the same statin, with the same 50% reduction in heart attacks. So will you be able to reduce heart attacks in that group by 50%?
No, because not everyone with high cholesterol will have a heart attack. OK, so how many people with high cholesterol will have a heart attack? It’s hard to say because that would depend on many other variables. But it’s probably not half (depending on how high their cholesterol is), and it may not be 25%. Around 75% of heart attacks occur in people with normal cholesterol.
So treating people based on one risk factor (say, high cholesterol), will result in you treating a lot of people who will never have heart disease. Whether you treat them or not.
Suppose you had a crystal ball, and you picked out 1,000 people with high cholesterol that you somehow knew would never have a heart attack. Your crystal ball said they were all going to die of pneumonias or car crashes or something. Ok, so now you give all 1,000 of them a statin. What will your success rate be? Zero percent. Because none of them were going to have a heart attack to begin with.
So let’s suppose, that in your high cholesterol study, 25% of the people in your study group will eventually have a heart attack. And suppose statins reduced the risk of heart attacks by half. In that case, the very best your statin could do would be to reduce the risk of heart attack by 12.5%. So you’re only getting a 12.5% reduction in heart attacks, even though the risk reduction from our hypothetical statin is 50%.
But it will probably be less than 12.5% because not everyone will take the medicine properly, some of them will take up smoking, some of them will get older during the study and develop diabetes, and so on. So in your high cholesterol study, a drug with risk reduction numbers of 50% may work less than 10% of the time.
Why? Because a lot of the people that you’re treating didn’t have heart disease to begin with. Whereas in the secondary prevention study, they all had heart disease.
In my speeches, I tell my doctors that if they’re using statins for primary prevention, they had better know what they’re doing with diagnostics, or else whatever treatment they use won’t work. No matter how effective it is. And if all they’re doing is giving statins to people with high cholesterol, well, good luck.
Ok. So after all that, let’s talk about masks.
Do masks prevent the spread of COVID-19? Maybe. Some studies have shown no benefit at all, but some have shown small risk reductions. Not 50%. Maybe 10%? If you choose your data carefully and squint a bit when you read the stats section, maybe 20%? Maybe. But probably less than that. And again, according to several studies, it’s close to zero percent. So the real benefit of masks, under ideal circumstances, is difficult to say for sure.
My point, though, is that it really doesn’t matter how effective masks actually are.
Let’s say for example that you went to a grocery store last year, at the height of the pandemic. Suppose there were 100 people in that grocery store. And suppose all 100 people, including you, are wearing masks. Ok.
How many of those people have active COVID-19 infections at that moment? Half? Of course not. All of these people are healthy enough to go shopping. 25%? Heavens no. Let’s just say for the sake of argument that there are five of those 100 people who have active COVID-19 and don’t know it. If they’re feverish and coughing their heads off, you’ll probably give them some space anyway. So I’m talking about asymptomatic, or nearly asymptomatic, but active, COVID-19 infection. Again, let’s say there are five of those walking around your grocery store.
So your risk of becoming infected with COVID-19 is 5%, then, right? Well, no. How many of those 100 people are you likely to have extended close contact with?
Remember that the vast majority of the infections we tracked occurred at home, or other places where people are in very close contact, in very small spaces, with so-so air circulation, for extended periods of time. Transmission of COVID-19 in public places is rare, whether you wear a mask or not.
So you could argue that we should be wearing masks at home, but not once you leave your house. I don’t think that would help much, either, but I suspect it would work better than wearing masks in public. That would be an interesting study. And good luck enforcing compliance in that study. Talk about hypothetical benefits…
Anyway. So those five people wandering around the grocery store do not give you a 5% risk of catching COVID. Let’s just guess your risk of catching COVID-19 in that store is 1%, just for the sake of argument. I think that’s really high. Some of you may think it’s a bit low. But whatever.
Remember that on that floating petri dish – The Diamond Princess – over 80% of the elderly passengers did not get COVID-19. And they were up to their necks in it for weeks. And they were all old. They weren’t middle-aged healthy people just wandering around a grocery store for 20 minutes.
But again, the exact number doesn’t matter much. Pick a number that makes you happy. But regardless, it’s low.
Ok.
Now, let’s suppose that masks can reduce the risk of COVID-19 transmission by 20% when used properly. Firstly, I think that’s extremely optimistic, based on the studies we have. And secondly, I would guess that the number of masks in that grocery store being used properly is vanishingly small. But let’s suppose that the studies are wrong, and the grocery shoppers are unusually vigilant with sterile techniques, and the masks in this grocery store actually do reduce the risk of COVID-19 transmission by 20%. Right.
So if your risk of getting COVID-19 in that store was 1%, and you reduce that by 20%, now your risk of getting COVID-19 is 0.8%.
So you reduced your risk of COVID-19 by 2 people out of 1,000. That’s not nothing. But it’s close. And again, I think that’s extremely optimistic.
Why are masks so ineffective, even if we presume that they are effective? Because almost everybody that is wearing them does not have COVID-19. So in those people, there is no chance that they could provide any benefit, regardless of how effective they are.
It’s like giving statins to people who have high cholesterol but don’t have heart disease.
Ok, look. The stats geeks out there are going to notice some shortcuts I took with my math. I used simple numbers simply to make a point. This isn’t meant to be a precise statistical analysis. And others may argue that it’s no big deal to wear a mask. With no real downside, why not, right? And the fans of masks may be plugging more optimistic numbers into my ham-fisted formulas above. But it won’t matter much. Maybe a little. Not much.
Also, just suppose that a new study came out tomorrow that showed that masks reduced the risk of COVID-19 by 50% or something, in that particular study, when you use a certain mask, with a certain sealant, for a certain period of time, or whatever.
Ok, that’s swell. But I don’t care.
Not until somebody can show me some very real benefit in an actual population of actual people leading their actual lives. If masks work only when used properly, but no one uses them properly, then that doesn’t help.
The research on statins was very exciting early on, but their use didn’t really take off until doctors started to see heart attack rates drop significantly in the population at large. That meant that they didn’t just work under the ideal circumstances of a controlled study – they worked in the real world, too. That’s when we really started using a lot of statins.
Also, you can reasonably argue that masks were a good idea, because they got people to leave their houses, and they got the economy going again. That may be true, but that’s not my point.
And I emphasize that I’m speaking of normal people wearing masks in public. I’m not discussing the use of masks or other protective equipment in medical settings. That’s not my point, either.
So what is my point?
If you feel safer wearing a mask, by all means, be my guest. I don’t think they’re entirely benign, but I agree that the risk from wearing masks appears to be small. So whatever.
But if you’re curious as to how much protection that mask is actually providing you, color me skeptical. I obviously don’t know the real benefit of wearing masks, but I strongly suspect that it is extremely, extremely small. A suspicion that has been borne out by population studies around the world.
I openly acknowledge that I don’t know the precise risk reduction provided by masks used in public settings. But I don’t think anyone else does, either. I’m just saying that whatever that number is, I suspect it’s really low.
And I think it’s important that we think this through because this is going to come up again. Probably soon. And next time, we should have a plan.
A plan with some real-world data behind it.
Our response to this virus has been a fiasco. I hope we learn from it.
I suspect we won’t.
Published in General
Wait I thought a bunch of studies in the 70s and 80s showed high colestorla had nothing to due with increased heart disease? Its one of the biggest myths doctors who have not been educated on think matters. Secondly there is only one study that shows mask work in public areas that I know of. They were using medical grade disposable mask. What am I missing. The other sevenish show no effect from what I remeber reading. Are their studies I have missed?
That every government claims moral authority for its existence and actions.
One takeaway I got was that any “study” or “poll” can be constructed to get the desired result. Even if the result actually doesn’t get the desired result, the data can be presented in such a fashion (e.g. lying with statistics) so as to give a false impression . . .
I beg to differ. Our country was founded on Judeo-Christain principles, not merely “inspired” by them. The Founding Fathers also made sure that natural law would be applied to everyone regardless of their faith (or lack thereof) . . .
In my view it is extremely difficult to lie with statistics if the person you are talking to actually understands statistics.
The average person doesn’t, which is why the book was written. The shuckster can even take advantage of the average person not understanding the differences between the mean, the median, and the mode.
It was founded on ideas from the Enlightenment with the assumption/expectation that people would live the principles and virtues of a religious people. In the same way that Adam Smith opined that free markets only work if there is a system of law that protected contract and property rights AND a culture that enforced and expected honorable behavior, I think that guys like Jefferson and Franklin (who were not deeply religious or orthodox practitioners) expected that the freedoms made possible by the Revolution needed to be exercised in ways guided by conventional Judeo-Christian morals. I think the Founders believed that if ever we sucked as a people, if we had no honor, no virtue nor moral ambition, then the whole project would go right into the crapper no matter how well it was structured.
In theory, if religion waned in the USA but some other belief system instilled the same necessary range of virtues and constructive behaviors, then the project would survive. In practice, I don’t know of such an alternative. Mushy, feel-good civic religion alternatives never struck me as a viable option.
How can we discuss statistics without first facing up to the fact that half of American kids are below average in math?
We never should have ended property restrictions with regard to voting rights.
The principal in my daughter’s elementary school once told me that it was his goal that all of the kids in his school test out above average. I told him, “Good luck.”
If thinks that his goal is ever obtainable, he is a fool. The evidence is depressing clear that you can’t make stupid people smart. I understand that no one wants this to be true but it is. We must endure the brutal empirical reality.
Man, I’m going to feel like the wet blanket here, for two or three reasons. Not exactly a novel situation for the least fun guy in the room, of course, but stil.
Okay, first, Doc’s comment was a joke, Henry: everyone can’t be above average, by the definition of average.
Secondly, of course the principle means above the general population average, not the average of the kids in the school, the latter being impossible because math. But most of us get the joke.
Thirdly, whatever the reality is regarding “making stupid people smart,” it seems likely that most normal school children can be taught math sufficiently well so as to perform reasonably well on standardized tests. That doesn’t require increasing inherent intelligence, merely competently teaching and drilling. If he were to impose reasonable classroom discipline, teach from a decent textbook that eschewed whatever Godforsaken nonlinear multimedia claptrap currently lards down elementary texts, and spend less time on fluff and more time on basic instruction, I’m sure Doc’s daughter’s principal could meet his goals.
Sorry to spoil a Lake Wobegon moment.
I agree that any person can improve their skills in math, English, science, etc… That is the whole purpose of schools. I never believed Charles Murray’s assumption that IQ is permanently fixed in people and cannot be improved upon. Every single time I have seen sample questions from IQ tests I have come to the conclusion that these types of questions can be answered more accurately with beforehand study.
My understanding — which could easily be incorrect — is that IQ is generally a pretty good proxy for a general level of cognitive function. I remember hearing someone, I think it was Murray, explaining that people who measure high on standardized IQ tests generally perform well across the board — are smart at most things, in other words. This contradicts the popular misconception that very smart people tend to be really good at one thing and correspondingly weak at lots of other things.
I don’t doubt that someone can study for an IQ test and perform well on it as a result. But most people don’t study, and serious testers probably administer multiple tests to filter out those who do.
What I always come back to in this kind of discussion, and the reason I tend to think our friend Henry C. is missing the boat here, is that we can achieve far more improvement in human performance through better teaching than we can through biology. I think we leave an awful lot of chips on the table when it comes to the way we teach our children, and we should work to fix that rather than waste time imagining dumb stuff like the gentle eugenics Henry C. keeps going on about.
It would not surprise me if the Principal actually said and/or believed in the quote. Thus my literal interpretation.
I entirely agree with you on improving our system of education through discipline and testable means of measuring knowledge and skills.
To address Steven Seward’s idea that, “I never believed Charles Murray’s assumption that IQ is permanently fixed in people and cannot be improved upon. Every single time I have seen sample questions from IQ tests I have come to the conclusion that these types of questions can be answered more accurately with beforehand study.”
That isn’t Charles Murray being weird. That is the consensus opinion of serious scholars after nearly a century of research. Additionally, nobody wants it to be true that I.Q. is static and everyone is interested in permanent increases to I.Q. No one has yet found a way.
You are correct that you can study up for an I.Q. test and you can get a higher score. But eventually, your I.Q. almost always goes back to what it was genetically.
Getting back to the original subject, I actually believe that mask mandates make the infection spread worse.
I base that on the overwhelming evidence that infection rates skyrocket in nearly every place in the world after mask mandates are put into effect. Here is one of many websites that show infection rates plotted on graphs along with mask mandates.
This would seem counter-intuitive even if you think masks do nothing. I have heard explanations by microbiologists on two factors that nobody seems to consider when wearing masks. The first is that most people don’t wear masks as tight-fitting as hospital staff are required to do. Your breath shoots out of the loose openings in a jet of air that travels much farther than if you just breath normally without a mask.
The second point has to do with the concept of stopping small water droplets with a mesh cloth. While it is admitted by scientists that the finest mesh in medical masks is not nearly fine enough to stop individual virus particles, it is assumed (and tested) that the masks stop the larger water droplets containing virus particles exhaled by the infected person. There is a big BUT in this assumption that nobody seems to think of. But what happens when the water droplet evaporates in the mask due to the wearer’s breathing?
A mask is not a black hole where things just disappear into nothingness. Those virus particles, now free of their moist anchors, are simply blown through the mask and into the atmosphere by the wearer. Without the mask, most water droplets fall harmlessly to the ground, but with the mask the virus suddenly becomes aerosolized.
He did say that. And you’re right, if his kids are really smart, maybe they can beat the average. But still, I find that possibility remote.
And I found it to be a silly thing to say.
The problem is that hypothesis testing is broken. The publication standard is for the p value to be less than 0.05 for a result to be declared significant. The Stat 101 explanation for this is roughly that p represents the probability that the result is wrong. But that’s not really accurate; moreover the 0.05 value was just plucked out of the air decades ago. It also holds a large number of assumptions about the study and the data etc. that cannot be guaranteed.
This is one of the drivers of the ‘replication crisis’ in scientific publishing. Now add dishonesty: with a huge data set and modern data mining techniques, researchers like Brian Wansink can play a game where they just endlessly play with the data until a p value under 0.05 shows up. Then you see if you can post hoc make up a hypothesis and publish as though you had not turned the process on its head (you’re supposed to start with a hypothesis to be tested).
Between the deep problems even with honest research and the potential for fraud, many professional societies and journals are thinking about new criteria. Some have proposed only allowing publication if p is below 0.005, while others say that the whole idea is shaky in its foundations and the old tests just don’t work.
Here is an amusing article on the p value problem; it contains links to other articles that have more details, for anybody brave or foolish enough to get down in the weeds of this issue.
Lake Woebegone Elementary was it? I suppose he meant national averages.
Pat Moynihan once joked that education scores got better as you moved to the Canadian border so we should just move all the kids from Mississippi to Michigan and Minnesota.
Detroit?
First, when you see cases going up, look to similar adjacent states/countries to see if they are going up at the same time. The northeast states (NY, NJ, MA, CT) have identical curves on identical dates. The mid Atlantic (PA, MD, DE and initially VA) peaked about 7-10 days later. (VA is now more like NC) . The midwest in the fall, the southern tier at the end of July. No alleged superspreader event (Thanksgiving, the Sturgis bike rally etc) made any curve go up. No policy intervention made it go down.
The point is COVID was always gonna COVID and spread and receded on its own terms in response to whatever conditions and targets it finds optimal.
We knew in advance that lockdowns, masks, quarantines, and border closings would have marginal effects at best but we pretended they were working and that the virus would have broken out and killed us all if restaurants allowed 51% occupancy or five or ten people went to the grocery store unmasked.
Yup. This page is a great resource for seeing that. Texas is big, and it has lots of different weather. The drop-down box at the top lets you look at any part of the state you want. So you can kind of tell right away that knowing the totals from the whole state has little to no predictive power for what’s going to happen next.
Here is Amarillo:
Here is El Paso:
Here is the Lower Rio Grande Valley:
And here is Dallas/Fort Worth:
You can look at many more, and they are very diverse in terms of the pattern over time.
It’s that bad, eh?
Steven,
Is the purpose of schooling only to teach skills, or does it also include teaching people to think?
I mention this controversy not to engage in dialog on it, but only as a reminder that some of us strongly disagree with your answer.
Notes:
It is a very important question for a society. Education cannot achieve its purpose if it doesn’t decide what its purpose is. Either answer leads to other questions.
Ricochet has adherents of both positions, and they quite properly debate it from time to time. Rarely does someone switch.
I’m not sure what you are disagreeing with. I am all for teaching students how to think, but I was originally responding to the question of whether or not someone can improve his IQ, not what should be taught in schools.
It seems pretty obvious that schools should, and perhaps must, do both.
There’s at least one more aspect to what schools may choose to inculcate, and that’s a set of values. I think determining the scope and particulars of that aspect of teaching — and deciding what compromises, if any, must be made elsewhere to fit that teaching in — is where the serious debate lies.
Teach primary students basic skills, provide them with essential instruction in basic problem solving, and do it in the context of a moral framework that is as close to universal for our culture as practical. That moral framework would include kindness, courtesy, thoughtfulness, honesty, helpfulness, responsibility, patience, respect — and a few other things. It needn’t, and shouldn’t, involve any instruction regarding sexual technique, birth control, men dressing as women, white guilt, etc.
That to “improve skills in math, English, science, etc.” is “the whole purpose of schools”.
Perhaps I did not elucidate all the possible purposes of public schooling in my sentence. Will you be keeping me after class, giving me demerits, or writing a note to my parents? Lol.
You did elucidate all of them according to the “skills” side of the dispute (all the purposes are subcategories of one purpose: teaching skills).
But it seems that maybe you didn’t really think that, but agreed all along that there are two overarching purposes (teaching skills plus teaching to think).
Anyway, sorry to hijack the thread, didn’t mean to.
Mask don’t work and are stupid.