undefined | Better HN

0 pointsFollowingTheDao3y ago0 comments

> If you manage to model the "noise" and make it somewhat deterministic then it's not noise any more.

Yes. That is called doing science.

> If genetic variations really are the reason for these variations and some people are indeed measurably harmed by these compounds then it would be a very interesting and somewhat alarming result, but that's not what the study says or what we can conclude from it.

What would make me not think that? Maybe we should investigate it. That is also called science. Saying anything is noise only disregards the response of that part of the sample as useless. It is not useless.

It is why some of us are more sensitive to to certain diets: https://www.chop.edu/conditions-diseases/carbohydrate-malabs...

0 comments

parineum3y ago

You know what isn't science? Not doing any of that and just going off on a study that's led to more questions (aka, science) in the comments of Hacker News.

Their study had a scope, they did the study and found some results then drew some conclusions. They also found the study wasn't large enough to draw all conclusions because of noise, something they didn't know before the study.

> Saying anything is noise only disregards...

Saying something is noise indicates that the data isn't clear enough to draw a conclusion and more specific and targeted studies are needed to draw said conclusion.

It certainly doesn't mean that they're ignoring the plight of... you.

FollowingTheDaoOP3y ago

I was commenting on how the original commenter was using the term noise. Not how scientists use it.

Listen, the term noise is probably the worst term for this data. Because we don’t know if it’s noise until we examine to see if it is noise. So until we know it’s noise we can’t call it noise. It’s like you’re walking into a crowded room and you’re trying to hear one thing but there is too much “noise”. This assumes we know what we’re looking for in the first place.

j / k navigate · click thread line to collapse