[ricardobeat]

The recent Bangladesh study with 350k people found a ~10% reduction in cases, from only 40% mask adoption.

The two-cities trial early in 2020 in Germany saw a ~20% reduction. Those are in the number of infections - the number looks low but can make a massive difference in the actual transmission rate. In the German case, after a month infections dropped close to zero.

[kyrra]

That study is flawed in many ways. And one thing to point out, if you do take the study at face value, cloth mask do virtually nothing, only surgical masks did something ( AKA the intervention group).

The study was not blind, so the intervention group (those that wore surgical masks and showed the large efficacy increase) we're paid money to be part of the group. The non-intervention groups (cloth and control) did not receive any monetary compensation.

Recording of masking was done via people observing mask wearing, which also could potentially mess with the data ( as this was not blind).

Reporting of covid symptoms was entirely based on the population manually reporting it, there was no random testing to see variances in covid.

Only around 30 to 40% of cases were actually verified via a test.

The education that came with the intervention group may have caused the older population to stop going out as much, which impacted the result.

My take is that the study proved you could pay people to mask, but it's efficacy result is dubious.

CATO has a meta analysis about mask studies that is interesting (pre print still): https://www.cato.org/working-paper/evidence-community-cloth-...

[ricardobeat]

I'm pretty sure the cloth masks debacle was settled sometime around Q1 2020, a year and a half ago. Anyone arguing about this needs to find a better hobby. They've been not recommended, or explicitly banned from airlines and other places, for a long time. They just confirmed this in a very rigorous controlled study.

> were paid money to be part of the group

That's simply not true. This is the document describing the intervention to raise mask usage: https://docs.google.com/document/d/1mgY6k5SooeMt6PIqwx-7z5LZ...

It says they tested monetary & non-monetary incentives, but if you look at the execution table, it's all "Public Reinforcement". The conclusion was that Nudges and incentives outside of the core NORM [1] intervention had no effect on mask-wearing..

> Recording of masking was done via people observing mask wearing

This is good. It means they observed the overall effect on the entire population. Some previous studies relied on self-reporting which is not as reliable.

> Only around 30 to 40% of cases were actually verified via a test.

You can't force people to take a test. But the rate of positives within the ones that agreed to collection was similar to the overall self-reported one. The study goes into this at length. There is a whole section trying out a different approach where they assign the average soropositivity to non-consenters, instead of excluding them, and that makes the results even stronger.

> The education that came with the intervention group may have caused the older population to stop going out as much, which impacted the result.

That sounds like a very random hypothesis. I can come up with another dozen of these. Maybe it rained more? Too cold? Maybe there was a soap opera on, that 60 year olds love to watch? You'd think a dozen scientists from Yale, Stanford, Berkeley, John Hopkins & others would find a way to control studies for external factors... if it was this easy to challenge results you could do it for basically every paper ever published.

The paper is available for free here: https://www.poverty-action.org/sites/default/files/publicati...

[1] no cost, offering, reinforcing, modelling

[kyrra]

Cloth Masks: I think you're wrong here? Per the CDC[0] (which all airlines seem to link to), it says:

> Cloth masks should be made with two or more layers of a breathable fabric that is tightly woven (i.e., fabrics that do not let light pass through when held up to a light source).

So cloth masks are allowed. Also, this is an issue that I agree has been settled for a while, but the media/political effort to push it has been minimal, it's always "mask up", without going into the details, which can be very important.

Intervention & Money: The doc you link just says they tried different approaches, but don't seem to details the differences with the different motivational types. It's not clear to me from what I saw that they really dived into this. And when there are any kind of rewards (ie: not blind), you will get different results in the intervention group that you were not expecting (people change behavior).

Observations of Mask Usage: I agree, this is probably the better than survey based (as from what I've seen, people self-report very differently than what they really do). This was a weak critique on my part and I would have to understand what alternatives there are that could be better.

Positivity Testing: I think my original point was moot here as well. I think the better argument here is that we do not know the change that covid had already spread in any given area prior to this test. I understand that they tried to group control and intervention groups that were near one another to try to cut down on this, but it is still a big blind spot for this study.

My other issue here is that there was no random testing done to find asymptomatic cases. This is a huge issue with this virus in general, and it makes our numbers not as good (The UK being one of the few countries that has this kind of data, but it's not truly randomized still).

Older Population Education: See this post on the topic[1]. The point she makes is that the reduction in covid by age group should have been equal if masks worked equally, but the results from the study show that the reduction in cases was mainly in the older age groups.

[0] https://www.cdc.gov/quarantine/masks/mask-travel-guidance.ht...

[1] https://twitter.com/Emily_Burns_V/status/1433122687765856259

[midasuni]

Cases in Scotland with mask laws and Hugh adherence seem higher than in England which doesn’t have laws and has low adherence.

Be interesting if anyone has done a proper comparison between say Glasgow and Manchester.

[ricardobeat]

I think I’ve heard that before, but how is it higher? From the usual dashboards Scotland has 130 cases per million, while England is at 150/million. The number of cases right now is also stable while the UK overall is going up?

[midasuni]

I last checked a couple of weeks ago when Scotland was on 370/100k and England on 350/100k, seemed to have flipped back the other way now

[ed_balls]

it's hard to compare cases without knowing how many tests were done and how they are administered.

[ralusek]

And where in a particular wave a given place is. And what the seroprevalence and other demographic factors are between the two populations.