Hacker News Reader

Comment by wvenable | Hacker News Reader

If the commenter read the study and found that they did not, in fact, account for that then it would be valid to point it out here.

Otherwise, it's just a waste of our time.

View on HN

By its fundamental design, an observational study cannot account for everything. That is the critique and it is a valid one.

Most observational studies do in fact account for the first thing a random HN poster can come up with 5 seconds after reading the title. So there is very little value in such a comment.

That's true, and it's also one of those factoids that everybody already knows but certain people can't help themselves from bringing up every time it's even tenuously relevant.

It might be common sense if you work with statistics, but most people don’t, and the only way they would know these things is if someone says them.

Regardless, it was the top comment so people clearly think it was valuable enough to vote for.

The fact that it was the top comment only shows that people agreed or liked it, not that it was actually valuable, accurate, or useful. Popularity and value aren't the same thing, correlation is not causation, etc, etc...

This was published in Nature, I'm pretty sure the reviewers understand and account for that. HN and reasonably educated people always think this is some huge gotcha, second only to "CORRELATION != CAUSATION!!!!"

Nature isn't an arbiter of truth, it's a publisher of papers. It's also meant to be read critically by people who see a paper like this and and think "interesting, but this is just an observational study, I'd like to see more research, maybe something with more controls before I believe this".

Say what you want about the "correlation doesn't imply causation", but many people do not understand why that is. As a shorthand for "hey, you don't know the statistics behind this, so don't take the pretty picture at face value", it's pretty good.

Reply on news.ycombinator.com