Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

The reproducibility crisis is present in almost all fields, though social sciences are definitely a place where it pops up often. Using it as an excuse to cut funding only makes the problem worse.

Computer science is also pretty embarrassing as a field of science: https://www.researchgate.net/publication/359147521_Reproduci... Based on this result (0%), New Zealand may as well dispose of all computer science studies as well.



> Using it as an excuse to cut funding only makes the problem worse

Or it's a wake up call to shift focus to better methodologies that will reproduce more reliably. Time will tell.

> Based on this result (0%), New Zealand may as well dispose of all computer science studies as well.

Lots of computer science research produces useful products that make their own money directly, and so the need for reproducibility is less if impact on the economy is a main criterion. This is not true of pure research in fields that only produce knowledge. Reproducibility is key to be confident you actually have knowledge.


> Lots of computer science research produces useful products that make their own money directly, and so the need for reproducibility is less if impact on the economy is a main criterion.

That research isn't being conducted by academics applying for pure research funding from the Marsden fund...


> The reproducibility crisis is present in almost all fields, though social sciences are definitely a place where it pops up often

Seeing this sentiment popping up a lot in this thread, but yet planes still fly, solar and batteries are improving like crazy...

We are seeing real tangible results that come from the output of other fields

We cannot really say the same of the softer sciences


Computer science is also pretty embarrassing as a field of science: https://www.researchgate.net/publication/359147521_Reproduci... Based on this result (0%), New Zealand may as well dispose of all computer science studies as well.

That paper investigates whether papers are well-formed such that the results can be reproducible. That's fair - many research papers today are vaguely written with insufficient information. However, that's not the issue at hand. The issue here is that, even with a perfect paper, social sciences are unable to produce replicable results.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: