5 ms·

That "one study in Bengladesh with masks" involved 342,183 participants and produced clear evidence that masks work. What is the point of more studies if folks

by seoaeu 5y ago

That "one study in Bengladesh with masks" involved 342,183 participants and produced clear evidence that masks work. What is the point of more studies if folks are just going to dismiss them out of hand?

mfcl 5y ago
That's not what I have been reading about the study. The evidence seems to be on the pro-mask side, but it does not appear to be _that_ clear. I'm not an expert in masks, health, or even statistics, so I'm not here to say you are right or wrong. I'm just pointing out that there is a debate. That itself is enough to want to look for more studies, don't you think? To answer your question, the point of more studies is to increase confidence in our opinions and decisions, make better findings, realize errors, etc. And you are quickly jumping to conclusions with "folks like you are just going to dismiss them out of hand". Now, my question would be: what do you mean by "folks like you"?
- seoaeu 5y ago
  Have you looked at the study itself? From the abstract: “Mask distribution and promotion was a scalable and effective method to reduce symptomatic SARS-CoV-2 infections.” The existence of “debate” doesn’t mean anything. Anti-maskers are always going to dispute studies like this, no matter how clear the evidence
  - opwieurposiu 5y ago
    Have you looked at the study itself? It found that cloth masks do nothing and surgical masks help a tiny bit. (control: 8.60%; treatment: = 7.63%) I suppose you could parse this as "masks work", I parse this as "if you are really worried about airborne virus, wear N95." Page 12: "no statistically significant effect for cloth masks" https://www.poverty-action.org/sites/default/files/publications/Mask_Second_Stage_Paper_20211108.pdf.pdf https://www.poverty-action.org/sites/default/files/publicati...
    - fxleach 5y ago
      Page 12 paragraph for context: "We find clear evidence that the intervention reduced symptoms: we estimate a reduction of 11.6% (aPR = 0.88 [0.83,0.93]; control: 8.60%; treatment: = 7.63%). Additionally, when we look separately by cloth and surgical masks, we find that the intervention led to a reduction in COVID-like symptoms under either mask type (p = 0.000 for surgical, p = 0.066 for cloth), but the effect size in surgical mask villages was 30-80% larger depending on the specification. In Table S9, we run the same specifications using the smaller sample used in our symptomatic seroprevalence regression (i.e. those who consented to give blood). In this sample we continue to find an effect overall and an effect for surgical masks, but see no statistically significant effect for cloth masks."
  - DenisM 5y ago
    > Have you looked at the study itself? Please don’t do this here. This is against the HN guidelines. https://news.ycombinator.com/newsguidelines.html https://news.ycombinator.com/newsguidelines.html Please don't comment on whether someone read an article. "Did you even read the article? It mentions that" can be shortened to "The article mentions that."
ghshephard 5y ago
Don't get me wrong - the Bangladesh study suggested that masks were a good public health initiative and should be used. What it didn't do was reassure me that they were a strong individual NPI. In much the same way that not masking was recommended early in the pandemic as a public health initiative to reserve the masks for first responders - wasn't particularly good advice for the individual.
- mrguyorama 5y ago
  The world has 8 billion people on it. Shouldn't we care about what helps everyone instead of what only helps you specifically? If masking helps everyone, then you should do it, even if it somehow magically didn't help you.
  - ghshephard 5y ago
    So - several threads here. It’s not unreasonable to try to understand what helps globally as well as individually. Early on, when supply chains for masks were stretched or non-existent, we were instructed not to wear masks. This was good advice globally, as it reserved the limited supply for those who could most benefit from them (first responders) - but bad advice individually. If you were looking for guidance to protect yourself and those you loved, following global advice was not the best path. Likewise, the current guidance that wearing masks - particularly if they are well fit FFP3/99%+ filtering, is good advice globally, but it’s unclear if it’s sufficient to reduce the risk down to <.1% for the individual (the point at which I would pretty much not worry about social distancing and avoiding stores, workplace, etc… I’m <65 and vaccinated, So I would be comfortable with a 1/1000 risk of catching Covid, particularly as I can usually avoid those scenarios) The question isn’t whether masks are an important NPI - they absolutely are - particularly if we could convince people to avoid going indoors unmasked and ensuring they are wearing (and not taking off) well fit FFP3s, the question is whether that is sufficient to mostly eliminate the risk of catching Covid. One lesson I learned here is that you really need to expand on your thoughts and not assume everyone else is already on the same wavelength as yourself.
rossdavidh 5y ago
It "produced clear evidence that masks work"...if you think a 10% reduction in the speed of transmission is "working". That's a pretty minimal return for a large intervention (30% increase in mask-wearing). Plus, they had to throw out 26 villages with 0 infections in order to get that much of an effect. Now there is a certain logic to that, but I wager that if most of those 26 villages had been ones with the intervention, they would not have thrown them out; I'm not aware of that part of the analysis being pre-registered. They probably don't have a non-zero impact, but it appears to be pretty small, even with a massive intervention like the one in Bangladesh.
michaelrpeskin 5y ago
Here's a breakdown of that Bangladesh study: https://www.skirsch.com/covid/BangladeshMaskReview.pdf https://www.skirsch.com/covid/BangladeshMaskReview.pdf
dang 5y ago
Please don't cross into personal attack when posting to HN. https://news.ycombinator.com/newsguidelines.html https://news.ycombinator.com/newsguidelines.html
- sdenton4 5y ago
  Lazy dismissal of scientific work is something of its own epidemic on HN these days. "Ah, but this study doesn't address this laundry list of things I just made up, and I haven't bothered to look at other sources, and I probably haven't bothered to even read this study, actually." It's especially common on COVID-related studies amongst the anti-maskers and anti-vaxxers on the site, but I have been seeing the same kind of lazy dismissal on lots of other posts as well. It's effectively an evolution of the Gish gallop. I find it both dreary and frustrating, personally, and it's not terribly surprising that people like the poster above call people out on it.
  - PpEY4fu85hkQpn 5y ago
    I agree. Every HN post is full of "well actually" non-experts who love dismissing the work of experts with a few sentences of layman talk.
  - dang 5y ago
    It's endemic on HN because it's endemic on the internet because it's endemic to humans. There's no "lately" about this. The question is, what do we do about it? Posting even worse comments, like the GP, is obviously not helping.
    - sdenton4 5y ago
      I like to believe that HN has at least been partially successful in holding itself to a higher standard than the internet at large, through a combination of community guidelines, overall community belief in the guidelines, and thoughtful moderation. You could, of course, say the same thing about personal attacks: They're endemic on the internet and to humanity, but we've mainly kept them out of HN using the above measures. Lazy dismissal is often the top comment on any given post, COVID or not. It's harmful to discussion and I would argue antithetical to the hacker spirit: If we only allow perfect airtight products/proofs, we miss out on discovery and exploration. Perhaps some guideline to avoid lazy dismissals (or something along those lines) would go a long way.
wskinner 5y ago
The Bangladesh study failed to produce a significant result when simpler nonparametric statistics were used to analyze the data. There were also significant discrepancies in the follow up rate between the control and treatment groups. And even if you ignore those issues, the absolute effect size (20 cases) was tiny. The effect size was expressed in terms of the relative rate of symptomatic seropositives, i.e., normalized by population denominators, which makes it appear larger and is sensitive to the follow up rate discrepancies (see above). Ben Recht at UC Berkeley has several good blog posts on this issue: https://www.argmin.net/2021/12/01/unblinding/ https://www.argmin.net/2021/12/01/unblinding/

by seoaeu 5y ago

mfcl 5y ago

That's not what I have been reading about the study. The evidence seems to be on the pro-mask side, but it does not appear to be _that_ clear. I'm not an expert in masks, health, or even statistics, so I'm not here to say you are right or wrong. I'm just pointing out that there is a debate. That itself is enough to want to look for more studies, don't you think? To answer your question, the point of more studies is to increase confidence in our opinions and decisions, make better findings, realize errors, etc. And you are quickly jumping to conclusions with "folks like you are just going to dismiss them out of hand". Now, my question would be: what do you mean by "folks like you"?

seoaeu 5y ago
Have you looked at the study itself? From the abstract: “Mask distribution and promotion was a scalable and effective method to reduce symptomatic SARS-CoV-2 infections.” The existence of “debate” doesn’t mean anything. Anti-maskers are always going to dispute studies like this, no matter how clear the evidence
- opwieurposiu 5y ago
  Have you looked at the study itself? It found that cloth masks do nothing and surgical masks help a tiny bit. (control: 8.60%; treatment: = 7.63%) I suppose you could parse this as "masks work", I parse this as "if you are really worried about airborne virus, wear N95." Page 12: "no statistically significant effect for cloth masks" https://www.poverty-action.org/sites/default/files/publications/Mask_Second_Stage_Paper_20211108.pdf.pdf https://www.poverty-action.org/sites/default/files/publicati...
  - fxleach 5y ago
    Page 12 paragraph for context: "We find clear evidence that the intervention reduced symptoms: we estimate a reduction of 11.6% (aPR = 0.88 [0.83,0.93]; control: 8.60%; treatment: = 7.63%). Additionally, when we look separately by cloth and surgical masks, we find that the intervention led to a reduction in COVID-like symptoms under either mask type (p = 0.000 for surgical, p = 0.066 for cloth), but the effect size in surgical mask villages was 30-80% larger depending on the specification. In Table S9, we run the same specifications using the smaller sample used in our symptomatic seroprevalence regression (i.e. those who consented to give blood). In this sample we continue to find an effect overall and an effect for surgical masks, but see no statistically significant effect for cloth masks."
- DenisM 5y ago
  > Have you looked at the study itself? Please don’t do this here. This is against the HN guidelines. https://news.ycombinator.com/newsguidelines.html https://news.ycombinator.com/newsguidelines.html Please don't comment on whether someone read an article. "Did you even read the article? It mentions that" can be shortened to "The article mentions that."

ghshephard 5y ago

Don't get me wrong - the Bangladesh study suggested that masks were a good public health initiative and should be used. What it didn't do was reassure me that they were a strong individual NPI. In much the same way that not masking was recommended early in the pandemic as a public health initiative to reserve the masks for first responders - wasn't particularly good advice for the individual.

mrguyorama 5y ago
The world has 8 billion people on it. Shouldn't we care about what helps everyone instead of what only helps you specifically? If masking helps everyone, then you should do it, even if it somehow magically didn't help you.
- ghshephard 5y ago
  So - several threads here. It’s not unreasonable to try to understand what helps globally as well as individually. Early on, when supply chains for masks were stretched or non-existent, we were instructed not to wear masks. This was good advice globally, as it reserved the limited supply for those who could most benefit from them (first responders) - but bad advice individually. If you were looking for guidance to protect yourself and those you loved, following global advice was not the best path. Likewise, the current guidance that wearing masks - particularly if they are well fit FFP3/99%+ filtering, is good advice globally, but it’s unclear if it’s sufficient to reduce the risk down to <.1% for the individual (the point at which I would pretty much not worry about social distancing and avoiding stores, workplace, etc… I’m <65 and vaccinated, So I would be comfortable with a 1/1000 risk of catching Covid, particularly as I can usually avoid those scenarios) The question isn’t whether masks are an important NPI - they absolutely are - particularly if we could convince people to avoid going indoors unmasked and ensuring they are wearing (and not taking off) well fit FFP3s, the question is whether that is sufficient to mostly eliminate the risk of catching Covid. One lesson I learned here is that you really need to expand on your thoughts and not assume everyone else is already on the same wavelength as yourself.

rossdavidh 5y ago

It "produced clear evidence that masks work"...if you think a 10% reduction in the speed of transmission is "working". That's a pretty minimal return for a large intervention (30% increase in mask-wearing). Plus, they had to throw out 26 villages with 0 infections in order to get that much of an effect. Now there is a certain logic to that, but I wager that if most of those 26 villages had been ones with the intervention, they would not have thrown them out; I'm not aware of that part of the analysis being pre-registered. They probably don't have a non-zero impact, but it appears to be pretty small, even with a massive intervention like the one in Bangladesh.

michaelrpeskin 5y ago

Here's a breakdown of that Bangladesh study: https://www.skirsch.com/covid/BangladeshMaskReview.pdf https://www.skirsch.com/covid/BangladeshMaskReview.pdf

dang 5y ago

Please don't cross into personal attack when posting to HN. https://news.ycombinator.com/newsguidelines.html https://news.ycombinator.com/newsguidelines.html

sdenton4 5y ago
Lazy dismissal of scientific work is something of its own epidemic on HN these days. "Ah, but this study doesn't address this laundry list of things I just made up, and I haven't bothered to look at other sources, and I probably haven't bothered to even read this study, actually." It's especially common on COVID-related studies amongst the anti-maskers and anti-vaxxers on the site, but I have been seeing the same kind of lazy dismissal on lots of other posts as well. It's effectively an evolution of the Gish gallop. I find it both dreary and frustrating, personally, and it's not terribly surprising that people like the poster above call people out on it.
- PpEY4fu85hkQpn 5y ago
  I agree. Every HN post is full of "well actually" non-experts who love dismissing the work of experts with a few sentences of layman talk.
- dang 5y ago
  It's endemic on HN because it's endemic on the internet because it's endemic to humans. There's no "lately" about this. The question is, what do we do about it? Posting even worse comments, like the GP, is obviously not helping.
  - sdenton4 5y ago
    I like to believe that HN has at least been partially successful in holding itself to a higher standard than the internet at large, through a combination of community guidelines, overall community belief in the guidelines, and thoughtful moderation. You could, of course, say the same thing about personal attacks: They're endemic on the internet and to humanity, but we've mainly kept them out of HN using the above measures. Lazy dismissal is often the top comment on any given post, COVID or not. It's harmful to discussion and I would argue antithetical to the hacker spirit: If we only allow perfect airtight products/proofs, we miss out on discovery and exploration. Perhaps some guideline to avoid lazy dismissals (or something along those lines) would go a long way.

wskinner 5y ago

The Bangladesh study failed to produce a significant result when simpler nonparametric statistics were used to analyze the data. There were also significant discrepancies in the follow up rate between the control and treatment groups. And even if you ignore those issues, the absolute effect size (20 cases) was tiny. The effect size was expressed in terms of the relative rate of symptomatic seropositives, i.e., normalized by population denominators, which makes it appear larger and is sensitive to the follow up rate discrepancies (see above). Ben Recht at UC Berkeley has several good blog posts on this issue: https://www.argmin.net/2021/12/01/unblinding/ https://www.argmin.net/2021/12/01/unblinding/