Hacker News Reader

Comment by cellar_door | Hacker News Reader

I don't know how you can take this data seriously. It's comparing population Vitamin D levels vs. population COVID mortality for n = 12 countries.

Are studies like this done to justify an actual Vitamin D study? They need to take a sample of confirmed COVID patients and test their Vitamin D levels individually.

View on HN

You don’t always need a big n to make causal conclusions. Do you know the story about the king who gets served a soup? Because he’s frightened that the soup might be poisoned, he gives it to his food tester first because the king reasons like this: „If my food tester (n=1) doesn’t die, then I probably won’t die, too, when I eat the soup“. Many great philosophers have argued, that people should use their common sense instead of requesting big tests. Therefore no big n‘s are needed. Start thinking.

??? The Kong’s conclusion wasn’t causal with any statistical certainty (which is what the above is arguing). It was anecdotal.

Maybe the poison doesn’t work for carriers of certain genes. Maybe it’s temperature reactive. Maybe it’ll take weeks to kill.

Just because the King observes something do not make his causal conclusions from that observation valid. Adequate sample-sized statistics aim to bridge that gap.

> The Kong’s conclusion wasn’t causal with any statistical certainty (which is what the above is arguing). It was anecdotal.

There ain't no such thing as statistical certainty my friend.

Well that settles it. Better keep on doing nothing!

(n=0)

http://gunshowcomic.com/831

it must suck to be the king, waiting a couple hours to a day to make sure the food tester doesn't die of a slow acting poison before eating oneself.

Just thinking.

I saw another one like this recently:

https://www.medrxiv.org/content/10.1101/2020.04.24.20075838v...

Started reading it and it has an incredibly small sample size. I agree, all these studies are weak. People are rushing to publish any crap they can right now.

Another study from Switzerland:

https://www.ncbi.nlm.nih.gov/pubmed/32397511 "5-Hydroxyvitamin D Concentrations Are Lower in Patients with Positive PCR for SARS-CoV-2"

Reply on news.ycombinator.com