[elcano]

The article makes you believe that the 11% benefit from using masks in the Bangladesh study was for the whole population (100%) using the mask (experimental group) vs the control group (0%) not using masks. However, the 11% benefit for an experimental group of only 42% of the them using masks vs 13% using masks in the control group. There was only a 29% point difference between groups (not 100%). And compliance was not 100% perfect (as usual).

Assuming that the study was correctly performed, maybe 100% usage difference between groups (all using masks in experimental groups vs nobody using them in control group) could have shown a much higher benefit. But that remains to be tested.

This article is, as an expert in the topic of agnotololy explains, an effort to "…spread doubt in the guise of balanced debate".

This other linked article is from early 2016 and is as relevant today as then: https://www.bbc.com/future/article/20160105-the-man-who-stud...

[nostrebored]

Wouldnt you expect an asymptotically increasing curve of utility?

I.e. if the jump from 10-40 is 11% the jump from 40-70 would be strictly leq 11? Haven’t thought about this intuition too deeply, fair warning

[elcano]

That's an intelligent hypothesis, but this could go one way or another.

I'd prefer to see empirical results.