[Tagbert]

Since most mask studies are done in high/extended exposure environments like hospitals, they are not the best model for most people’s use.

This summer, the state of Kansas did an (inadvertent) infield experiment where some counties implemented mask mandates and other counties did not. After a few week, the non-mask mandate counties infection rates increased rapidly. The mask mandate counties’ infection rate increased on slightly.

[dijksterhuis]

Link to a video presenting the findings of a study from the University of Kansas on this "inadvertent infield experiment"

https://mediahub.ku.edu/media/Masks/1_49bb9aid?campaign_id=1...

[irq11]

You should look more deeply into that Kansas study. They cherry-picked their data in a way that is extremely dubious:

https://mobile.twitter.com/ianmsc/status/1332407421151576064...

Turns out that if you pick slightly different start and end dates, you can get dramatically different results.

Even if you think there’s a justification for it, plotting the counties’ data next to each other shows that they’re torturing the data. That is a perfect example of the horrible, flawed methodology that has been getting published all year.

[lern_too_spel]

The mechanism by which masks would reduce transmission has been repeatedly studied in a laboratory environment, so in the absence of a properly designed randomized control trial, the prior should be that it works. Instead, in this thread, you have cherry picked tortured data and used flawed methodologies to say that they don't work. Saying that there are similar poor studies on the other side does not update my prior.

[irq11]

I have “cherry picked“ nothing. That’s the data from the study. The researchers picked the start date, arbitrarily, to produce the outcome they liked. It’s obvious.

By the way: there was a randomized controlled trial for masks and covid published just recently. They were able to detect no statistically significant protective effect:

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

[lern_too_spel]

I was referring to other studies that cherry picked for your position you posted "in this thread," not that study that supposedly cherry picked for the pro-mask position.

The study you just linked to is an excellent example of using poor methodology to support your position. If a study is not powered to detect the expected change (for the mask wearer, which is expected to be very small) with full compliance, I would certainly not expect to get a statistically significant result with partial compliance. It did not attempt to measure the effect on transmission from the mask wearer to others at all.

[irq11]

I “cherry picked“ no other studies for this thread.

The DAMASK study is literally the only RCT ever run for masks and Covid. You may not like it or what it says, but its much higher quality than any of these uncontrolled, correlative, pro-mask “papers” that have been rushed out since March.

> It did not attempt to measure the effect on transmission from the mask wearer to others at all.

Because you can’t. Short of locking a group of people in a room with infected people and seeing who gets it, thats an unfalsifiable hypothesis.

Fortunately, we know that masks are not magically effective in one direction only, and if they don’t work in one direction....

[lern_too_spel]

> I “cherry picked“ no other studies for this thread.

Here: https://news.ycombinator.com/item?id=25680698

> You may not like it or what it says, but its much higher quality than any of these uncontrolled, correlative, pro-mask “papers” that have been rushed out since March.

I showed you why it is not.

> Because you can’t.

And here you admit as much — that this study is useless to measure the effect of a mask mandate.

> Short of locking a group of people in a room with infected people and seeing who gets it, thats an unfalsifiable hypothesis.

And here you admit that it is impossible to do a study that proves the effectiveness of masks to the degree that would satisfy you. In the absence of such studies, we have to rely on our priors, which are based on laboratory-proven effectiveness of masks in reducing the distance and the mass of virus particle spread from the wearer combined with what we know about germ theory.

> if they don’t work in one direction....

You somehow ignored that your study does not show that it does not work in either direction. It was not designed to measure effectiveness in one direction, and it was not powered to measure effectiveness in the other direction.