whomever you are quoting on Bangladesh study raises a couple of interesting questions based on the fact taht a study like that is not blinded – people know that they are getting masks. But they skew the results somewhat – 7% number is for cloth masks, higher for surgical masks – and when compliance is only 40% indeed.
Biggest objection seems to be – the commenter can not envision what is a mechanism of different rate of decreasing symptoms between older and younger.
Effects on older people were indeed more significant. Was it partially due to increased distancing (as reported)? possibly. I don’t see anything bad in that. The whole goal of the study was to find public policies that work. Remember early incorrect predictions by virologists playing public policy experts saying that masks will cause decrease in SD? That seems to be wrong.
But a simple explanation that removes the objection is that effect of the doze does not have to be proportional. An older person might get such more significantly with more exposure and masking mitigates those cases.
Another note: mask compliance in “blue America” is well south of 100%.
Again, the objections raise a couple of interesting points. I would try blinding it, for example, by providing defective surgical masks as controls. But a claim that “fancier” masks lead to consistently different rate of reporting is grasping for straws. To the strength of the study, there are multiple other intervention that they tried that did not show any effect. So, there is no easy bias introduced into this data.