Hacker News new | ask | show | jobs
by card_zero 552 days ago
And then you find it was wrong, and you keep it, and make protestations.
3 comments

they found it was invalid in the short term, for this particular study. the long term is still an open question. which is why they’re pointing that out.

saying “we thought this would happen, it didn’t, but maybe there’s just something to do with our study that meant we disnt see the result that confirms our hypothesis” is a perfectly valid conclusion.

I’ve seen the same thing in many A/B tests
I’d humbly suggest “protestation” isn’t the right word here