"This doesn't look like quite such good advice now that the Bangladesh study is out (https://www.poverty-action.org/sites/default/files/publications/Mask_Second_Stage_Paper_20211108.pdf.pdf)."

Yea....I'm not sure I was giving "solid advice" as much as I was just saying "masks aren't a magic supersuit, and if they're not a magic supersuit, stop getting all agitated about them."

That, and pointing out that data from a lab is not data in real life.

As far as the Bangladeshi study is concerned, you and I came out of it with almost the exact same impressions. The data was very noisy, but I remember thinking that masks probably decreased your chances of getting covid by about 10% or so on average, which makes them out to be a little less effective than I thought when I wrote the first article. Then, I would have SWAGed a 20% guesstimate.

And I agree, even 10% is not nothing.

"And if you might end up accidentally causing me serious harm or death, I don't think it's terribly unreasonable for me to want you to take a pretty low-effort action that might reduce that by ~20%, is it?"

No, but this is one of those things that goes to individual perception of risk, and if you're a student of that topic at all, you know that humans are pretty terrible at assessing risk. As the Freakonomics guys would have put it, paraphrased, "if you really cared about the safety of your kids, you'd fill in that backyard swimming pool." :-)

"(However, I agree that the focus on "wear a mask, any mask" rather than "let's carefully evaluate the effectiveness of different masks" is exactly not following the science...heck, we hardly even managed to get funding together to properly do the science.)"

Well, yea. A bandana is not a mask, really, but neither are those Chinese-made "not for medical use" masks with the ear loops that have been heavily used for the last 20 months, really.

Obviously, N95's are going to give you different (and better) results, but there's way too many people who can't afford them.