This line about mask usage in Melbourne from a Burnet Institute researchers' opinion piece ( https://www.smh.com.au/national/at-a-delicate-stage-how-to-crush-the-threat-of-a-second-covid-wave-in-sydney-20201226-p56q89.html)

Looks to be based solely on ... counting the number of photos of masked Melburnians printed in The Age? ( https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3714648)

dunno about that methodology

H/T several people on twitter who shared this preprint

Also, this is a pre-print that has not been peer-reviewed. If we're going to be publishing findings from preprint papers in major news outlets it would be a good idea to .... follow the recommendations of the journal

If I were inventing a novel methodology for estimating a population parameter, I would probably include some discussion of the method's usefulness in my paper that I am submitting to one of the world's most well-known medical journals.

But I'm not a researcher.

There was also a survey where people self-reported how often they wore a mask, with data presented from a few days either side of them being presented (a shorter window than the newspaper photograph review)

It... doesn't look quite as impressive.

The preprint also says on the fourth day masks were mandatory '100% reported “always” doing so'.

That result comes from a sample size of... seven (not a typo)