Hacker News Reader

Comment by irq11 | Hacker News Reader

The open letter for face masks is a great example for why this kind of thing is dubious.

All prior high-quality evidence disputed the claim:

https://www.cidrap.umn.edu/covid-19/podcasts-webinars/specia...

The subsequent evidence of real-world effectiveness isn’t great: cases are skyrocketing around the northern hemisphere, and mask mandates aren’t making a difference. The only randomized controlled trial ever conducted for masks and covid was not able to detect any protective effect, yet was rejected from multiple journals because of “consensus”:

https://www.medpagetoday.com/blogs/vinay-prasad/89778

Science by signature count isn’t science; it’s a popularity contest. It’s perfectly fine to advocate for a belief, but nobody should pretend that they have the “scientific” argument simply because their signature list is longer.

View on HN

> cases are skyrocketing

And comments like yours are one of the reasons.

The study shows that masks do little to protect the wearer - which is kind of obvious as it’s not going to filter incoming air very efficiently, that has never been the purpose of surgical masks. They don’t protect against airborne pathogens and every medical professional knows it.

You need a crítica mass for masks to start reducing transmission rates. Mask use is nowhere near what it should have been, so you cannot take any conclusions out of that.

> Mask use is nowhere near what it should have been

All available evidence shows that you’re wrong.

Self-reported mask usage rates are over 90%, nationwide. Moreover, many of the places with the highest compliance rates (e.g. Los Angeles) have the fastest-growing case counts:

https://delphi.cmu.edu/covidcast/?date=20201018&signalType=v...

https://www.google.com/amp/s/www.washingtonpost.com/business...

https://www.cdc.gov/mmwr/volumes/69/wr/mm6943e4.htm

Because population density and culture are the other factors. In most cases, combined with limiting large groups and reducing the frequency of meetings, it is doing what it is intended to do - reduce the R0 factor. Even in LA.

Every Internet pundit who can take one number and divide by another, announces they have proven something or other. Give it a rest.

There’s always an excuse, but there’s never any data backing your claims.

Provide even a single shred of evidence that masks affect the R0 of the virus in the real world.

Modeling studies and correlations observed in April don’t count: most of the papers published this year looked at areas where case counts subsequently skyrocketed in the fall. My personal favorite is the mask study from Jena, Germany in the spring, which only made it into a journal right as cases were reaching an all-time high in the same area:

https://www.pnas.org/content/early/2020/12/02/2015954117

https://gesundheit.jena.de/en/coronavirus

From the first link you mention:

> we conclude that 20 d after becoming mandatory face masks have reduced the number of new infections by around 45%

A later increase in infections does not invalidate statistical significance for earlier studies. It’s just math, not “an excuse”. You’ll never see the data for how many cases would have been recorded without the measures in place.

You can see the full paper here: http://ftp.iza.org/dp13319.pdf

The second link you posted earlier also says conclusively that higher reported mask usage correlated with lower reports of symptoms. Are you even reading your own sources?

Or maybe you just have very skewed expectations of the impact? The Jena study estimates a 0.23 reduction in R rate. It is nowhere close to eliminating the disease, but this small reduction in transmission rate has an outsized effect on the total number of cases, hospitalizations and deaths, and that is what really matters (roughly, a 50% increase in transmission would result in 10x more deaths).

The inconvenience of wearing a mask is an insignificant price to pay even if the numbers are off by an order of magnitude.

> A later increase in infections does not invalidate statistical significance for earlier studies. It’s just math, not “an excuse”.

So your theory is that masks worked then, but stopped working now?

Convenient!

> You’ll never see the data for how many cases would have been recorded without the measures in place.

Well, yes, exactly...these are uncontrolled studies. Which is why they’re such garbage.

You can see that in the title of the paper you posted: “a synthetic control study”. That means they didn’t have a control. Which means all they can do is squint at correlations and try to claim they prove cause.

(oh also: the WaPo “data” is from a phone survey...so yeah: people who say they wear masks a lot claim they have fewer “symptoms”. Whatever that means.)

> So your theory is that masks worked then, but stopped working now?

This doesn't follow. It's completely believable that masks have continued to be effective at curbing spread by 40% continuously. However increased holiday travel, as well as colder weather and people hanging out more in doors caused an increase in spread that wasn't present in the summer.

So let me summarize:

Cases go down after mask mandate: proof that masks work!

Cases go up after mask mandate: masks still work!

This isn’t science.

You're right, your inference method isn't science. Science is "given two relatively similar municipalities, if one enforces a mask mandate, cases decrease in that municipality compared to the other". A finding that was repeatedly verified on the scale of cities, counties, and nations.

That months after mask mandates were enforced, case rates went up everywhere, including in places where masks weren't enforced, doesn't sound like it has particularly much to do with masks.

What isn't science, is every internet genius that can divide one number by another, or compare two numbers but ignore everything about the context, is suddenly a self-proclaimed expert.

Here's a notion. Listen to real experts, and do what they advise. Instead of spinning yourself up and posting nonsense.

No, the one from Jena was a controlled study, and they tried to exclude any biases like neighboring cities. It’s “synthetic” because you can’t create two equal cities and population to test, unfortunately.

Again, did you actually open and read the link? It’s pointless to discuss your feelings in place of the studies. You’re dismissing all of the data in that report based on.. what? Again, the increase in total cases tells you nothing about the efficacy of the measures. The controlled study does that.

Studies like these can be invalidated by further studies, or systematic reviews. They are not invalidated by popular feelings over news reports. Show me one of those statistically-backed rebuttals and I’ll believe you.

(The WaPo link was posted by yourself, most of those are self-reported. It’s good that you notice because it also weakens your own argument if you think about it)

It’s a “synthetic control” because they didn’t actually have a control group. So they tried to pretend their was such a group by taking data from elsewhere.

Yes, I read the paper.

How else do you propose a control group for an entire city, other than looking at other cities in the same region? Have you done research like this before?

Here you go: https://www.nature.com/articles/d41586-020-02801-8

Read that, chase down all the study links therein, and come back a week after you've found out you're wrong.

You believe self reported data? Self reported data is among the least reliable forms of evidence there is.

Can you point me to any data showing that random quality cloth masks even reduce coronavirus transmission?

The recent Kaiser San Jose case[0] where one worker infected 44+ fellow staff, all of whom were wearing masks, ... from inside a tree costume would seem to be a great counterexample.

[0] https://www.mercurynews.com/2021/01/07/kaiser-san-jose-fined...

Essentially everyone I see in the wild has been wearing masks for months, with almost perfect compliance since the Fall surge, including myself. But it seems to be mostly theater. It's like stopping BBs with a volleyball net.

It also ignores the fact that the eyes are a bigger attack surface than the nose or a closed mouth.

If we had stressed the importance of not touching one's face combined with aggressive sanitization of hands, faces, and surfaces, Covid-19 might have waned into a quiet death even without a vaccine, like SARS and MERS before it.

> If we had stressed the importance of not touching one's face combined with aggressive sanitization of hands, faces, and surfaces, Covid-19

This...was stressed, significantly. Do you not remember that?

Look at any 20 articles about Covid transmission from February to April. You'd be lucky to find one that stressed not touching surfaces over mask wearing.

In 37 California articles I saved from that period, exactly zero even mentioned not touching surfaces in equal or higher importance to wearing masks... and only 8 even mentioned surfaces as a transmission vector (generally, only doorhandles).

> In 37 California articles I saved from that period, exactly zero even mentioned not touching surfaces in equal or higher importance to wearing masks...

That's because surfaces aren't a major transmission vector. There was significant early concern about surface transmission, with recommendations to leave packages to quarantine for a period or wipe things with alcohol to sanitize them. There was a period where I'd wipe my phone and credit cards with alcohol everytime I reentered my house.

Pretty quickly though, the data showed that didn't matter, and so people (reasonably) stopped caring as much, instead prioritizing airborne particle transmission, which remains the main transmission vector.

> It also ignores the fact that the eyes are a bigger attack surface than the nose or a closed mouth.

Again, the primary purpose of wearing a mask is not to protect you but everybody else. By making it harder for you to spread your viruses. So, it only protects you if everybody else around you wears them.

Nevertheless, masks have a small effect on protecting with regards to viruses entering your mouth and nose, but none on your eyes, obviously.

Can you point me to a study showing the effectiveness of not touching one's face?

Or parachutes?

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC300808/

Literally everyone thinks their favorite pseudoscientific theory is a parachute.

Calling something a parachute is not science. It’s just a pseudo-intellectual defense of superstition.

I don't understand your hard-lined conclusion when the paper that the blog post references can't reach a conclusion either.

The blog post also reaches a semi-hard lined conclusion yet still leaves room for additional work as it should.

Problems with the study:

- Not everyone the participants are interacting with also have a mask (Blog post references this with "protects you, not me")

- The result is "Although the difference observed was not statistically significant, the 95% CIs are compatible with a 46% reduction to a 23% increase in infection."

- "The recommendation to wear surgical masks to supplement other public health measures did not reduce the SARS-CoV-2 infection rate among wearers by more than 50% in a community with modest infection rates, some degree of social distancing, and uncommon general mask use."

My point is "Masks are in effective" can't be concluded from this. It failed to show a reduction of 50%, sure, but what about 25% at a higher CI? Was it literally 0% at 5 sigma? Clearly more work needs to be done to show 0% at any meaningful CI.

If you don't believe in masks fine, don't go see other people - it's way more effective anyway.

You’re right, the Danish mask study showed that masks are consistent with every outcome ranging from an increase in infection to a reduction.

You spins the wheel, you takes your chances, I guess...

There's a difference between "mask mandates aren't making a difference" and "masks aren't effective." Take a look at Japan, where the government has done fuck all to deal with the virus, but people wear their masks.

> Take a look at Japan

OK: ” A record 7,841 new coronavirus cases were reported in Japan as of 8 p.m. on Friday. The nationwide daily tally has set a record for a fourth day in a row.”

https://www3.nhk.or.jp/nhkworld/en/news/20210108_40/

That’s some fantastic evidence that masks don’t work.

Le sigh. Just because infections are increasing does not mean masks don't work. Tell me, what would the numbers be without masks?

https://medium.com/incerto/the-masks-masquerade-7de897b517b7

Science is built on consensus

Also the fact that mask mandates are poorly enforced, or that masks arent a replacement for other forms of prevention, or that some materials dont offer the same level of prevention dont change the fact that masks work

> Science is built on consensus

That's definitely the messaging public representatives use when communicating science issues to the general public. It's very effective at getting compliance across a society where our instinct is to look to our peers to figure out how to behave.

As others have said, what has made science effective at being a vehicle of growth and change is that it is empirical / evidence based. New ideas slowly diffuse and get accepted as evidence grows for their support (which can take a long while).

Decision making by evidence is also pretty much the opposite of natural human group behavior. I recommend reading/listening to some of the newer research on human psychology of groups. It pretty concretely shows that people consistently choose group harmony + stability over converging on the truth / correctness. Basically in a group we value getting along with each other a lot more (being in agreement / having consensus) than we value knowing what really is happening. Also, scientists being humans are not immune to this behavior either which is why evidence based progress is slow and can be resisted by established groups for lengthy periods.

It is not. Science is “built on” evidence.

I am a professional scientist. Nobody has ever given me a pass on publication simply because they want to believe me.

That said, science is a human institution, and all humans have cognitive biases that can be exploited - particularly in large groups.

That's a terribly optimistic view of science, speaking as someone who was also in the same boat. They might not just accept anything, but it'll be much easier if they want to believe you. This goes quintuple for social science.

Yes, we agree on that. Which is, again, why we don’t do “science“ by length of signature list.

> Science is built on consensus

No it's not. It's built on hypothesis and experiments (empiricism).

Just because the nazi party banned "jewish" physics doesn't mean that "jewish" science wasn't science.

You think science is a democracy? If 51% of scientists say the earth is flat, then it's flat? You think each nation has their own version of science built on consense of their own scientists?

Reply on news.ycombinator.com