Hacker News new | ask | show | jobs
by justinjlynn 1503 days ago
> Why did the word 'risk' disappear here? I don't think that's a good null hypothesis at all.

What null hypothesis would you prefer?

> I am not saying that.

Then what are you saying?

1 comments

> What null hypothesis would you prefer?

"It has no lasting effects", probably.

And you need to make sure you're looking for both positive and negative effects.

"It's harmful" is a huge bias for a null hypothesis.

> Then what are you saying?

That you shouldn't assume it is harmful.

I didn't say anything about what parents should be "able to do". I didn't even give an opinion on the candy thing. I just think your justification is a big overreach.

But also a null hypothesis implies you're currently doing the testing, so that's a different scenario too...