News for hackers. Stuff that matters.

A comment on When physics meets biology: a less known Feynman. Parent comment

Re:When physics meets biology: a less known Feynman by nonbel on Tuesday May 15, 2018 @09:32AM (#17073763)
Not sure what you mean but my post was a sarcastic comment about how people choose an arbitrary level of significance so that the "right number" of "discoveries" can be published each year.

If there is a lot of data they would have too many "discoveries" (it would seem too easy) so they need a more stringent threshold. If the data is very expensive they would have too few "discoveries" so the threshold gets relaxed.

[ Parent | on HN ]
Re:When physics meets biology: a less known Feynman by vibrio on Tuesday May 15, 2018 @03:03PM (#17076712)
Pardon, I have a weak ability to detect sarcasm before finishing my first coffee. I fully agree that p-value limits are arbitrary, and what one is examining can greatly impact the interpretation of 'significance' or not. As you imply, data can be inflationary , incrementally increasing in volume and decreasing in value.
[ Parent | on HN ]