Hacker News Reader

Hacker News

Comment by thedrbrian | Hacker News Reader

Parent Full thread

thedrbrian

>parachutes

Very easy and simple demonstration though. With a high confidence value.

Whether masks make a 1% or 0.07653 difference or whatever is very difficult.

View on HN

It’s very tough to do a good double-blind parachute RCT.

https://www.bmj.com/content/363/bmj.k5094

....which, again, you don't need to do, because the effect size of parachutes is enormous. You must do them when the effect size is measured in single-digit percent (or less, as in the case of mask mandates).

You're not making a responsive argument to the parent.

how do you get an "effect size" or "confidence interval" without a properly controlled trial? I could claim that the "effect size" of umbrellas leading to rain is absolutely enormous, billions of gallons of it drop from the sky! You parachute believers are really something else.

The same reason that the argument is parroted by unserious people in "arguments" like the one above: some things are obvious. I don't need an RCT of dynamite to know that it is deadly.

If we do not allow exceptions for the obvious, our intellectual rigidity becomes a liability.

Exceptions for "obvious" stuff should be temporary. If a thing is so obvious and relevant to your life, then there should be mountains of anecdotal and scientific evidence proving the idea. Remember, lots of ideas seemed obvious at one time or another and turned out to be disastrously wrong.

I have it on good authority from the mask-questioning community that double-blind RCTs are the only valid tests. We should also prove Koch's postulates on aeroplane falls, just to make sure.

Your comment is amusing and biting! Sadly, with and irony are not particularly valued on HN. Direct clear expressions are preferable for the expected neurotypes of this pseudo-community.

If I had to choose, I'd choose a bias toward double-blind RCTs than the parade of obviously biased nonsense used to justify mask mandates.

Pseudoscience and sloppy thinking have a well-established constituency, and need no additional support.

Reply on news.ycombinator.com