Over the weekend, everyone's been talking about this new study on masks and COVID-19 that apparently shows that masks prevent the disease and that pretty much all transmission is airborne
As has been pointed out, the study is terrible and proves nothing 1/n
As has been pointed out, the study is terrible and proves nothing 1/n
2/n a couple of very nice threads looking at the issue:
3/n On top of these, I wanted to note a couple of other things that make this study obviously wrong in terms of design and conclusions
4/n The study is here pnas.org/content/early/…
Basically, the authors compared the rate of infections before/after mask interventions in three places to a linear trend extrapolation, and found that after masks there were fewer infections than before
Basically, the authors compared the rate of infections before/after mask interventions in three places to a linear trend extrapolation, and found that after masks there were fewer infections than before
5/n They also plotted linear regressions of pre-mask and post-mask interventions, showing a linear trend upwards and then downwards
Thus, masks caused the change!
Thus, masks caused the change!
6/n Now, this is at face value implausible. The authors have picked out masks as an intervention, but the reality here is that there was NO SINGLE INTERVENTION that you could easily point to at any moment in time
7/n In New York, in the same week that the governor signed a mask law, there were at least 2 other laws implemented regarding COVID-19, and many changes in policy as well
It's not as simple as a single intervention!
It's not as simple as a single intervention!
8/n Moreover, the authors have cherry-picked 3 specific places that appear to support their argument
This is obviously bad science - what about the rest of the world?
This is obviously bad science - what about the rest of the world?
9/n What about Australia, where we've never had mask laws (and most people don't wear masks)?
10/n Or South Korea, where masks have always been a major part of the pandemic response?
Both of these are direct counter-examples to the arguments made in the piece, but are excluded
Both of these are direct counter-examples to the arguments made in the piece, but are excluded
11/n There are dozens more such examples. While in some countries masking was temporally related to declines in case numbers, this is by no means true of every place in the world
12/n So the authors have cherry-picked their interventions, cherry-picked their locations, and as others have identified done quite a few other questionable things as well
It is ~very~ frustrating to see
It is ~very~ frustrating to see
13/n Especially since I am not against masks!
The point here is not that masks do or do not work, it is that this paper provides absolutely no evidence either way and is largely useless at determining cause and effect
The point here is not that masks do or do not work, it is that this paper provides absolutely no evidence either way and is largely useless at determining cause and effect
