Writing comics about the economy is not the road to immoderate wealth that one might suppose, which is why I have a day job writing about medicine. So when I started reading about how vitamin D can help with COVID-19, I wound up making this:
(This was made in early 2021; now it’s late 2021, and it holds up pretty well. This meta-analysis [https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8557713/pdf/12937_2021_Article_744.pdf] says that there’s no statistical effect, but it includes four sub-analyses, each of which are right on the statistical line–clearly there is an effect, and they just happened to cut the data in a way that juuuuust failed to show it four times. Also, that study confuses odds ratio and relative risk, so I’m not sold on their statistical chops.)
Leave a Reply