The new CDC study on mask effectiveness compares Kansas counties with and without mask mandates and finds that masks are effective.

Are they, really? (1/x)

https://www.cdc.gov/mmwr/volumes/69/wr/mm6947e2.htm https://twitter.com/ianmSC/status/1330595974394249216

The first red flag comes from the data they used for analyses. They pick two seemingly random weeks from before and after the mandates and compare the changes in case numbers. They find that cases doubled in non-mandate counties and slightly declined in mandate counties. (2/x)

While this looks like cherry-picking, the results are actually similar when we compare mandate vs. non-mandate counties in terms of overall increase since August. On average, cases per 100K increased by x19 in non-mask counties vs. x9 in mask counties. (3/x)

This got me excited for a minute. I actually thought masks may have made a difference. Wouldn't it be wonderful if we could actually eliminate a virus by just covering our mouths for a few months? Sounds too good to be true! Because it is... (4/x)

There is a huge elephant in the room: Urban vs. rural is a giant covariate. The most populated five counties are mask counties, causing all kinds of biases in the analyses. A fair comparison would be between counties with similar populations, which is what we do next... (5/x)

For this analysis, we pick 22 (11 mask and 11 non-mask) counties. See below for which counties are included. Their populations are almost the same, with average being 30K for non-mask and 31K for mask. (6/x)

Comparing their case increases from August 11th (the date CDC report used as the cut-off) till now, we get this: Counties without mask mandates saw their cases increase x5.77, while counties with mask mandates had an 8.37x increase in cases. (7/x)

I want to clarify my position once more: I am not strongly against masks. I would love for them to work and actually save lives. I am just not seeing the evidence... (8/8) https://twitter.com/covidtweets/status/1329260509363916800?s=20