Are other fields really better than psychology in that respect?
I failed to reproduce published work everytime I tried. In one case I was only able to reproduce results by using carefully selected starting parameters I received privately from the author; in another case the author just told me that the algorithm was very unstable and required a lot of tweaking (not mentioned in the paper) to work, and another author didn't tell me about the boundary conditions used in the paper that were left out because they were too busy. Those were papers from physics & computer graphics.
The cynic inside me assumes that non-reproducibility is pretty much the default for most published papers.
Comments
Reminder that psychology is in a reproducibility crisis; with ~30%-50% of studies being unable to be reproduced.
https://www.theatlantic.com/science/archive/2018/11/psycholo...
In which case, it's fairly safe to say... you're not that much worse off by following your gut and flipping a coin.
Are other fields really better than psychology in that respect?
I failed to reproduce published work everytime I tried. In one case I was only able to reproduce results by using carefully selected starting parameters I received privately from the author; in another case the author just told me that the algorithm was very unstable and required a lot of tweaking (not mentioned in the paper) to work, and another author didn't tell me about the boundary conditions used in the paper that were left out because they were too busy. Those were papers from physics & computer graphics.
The cynic inside me assumes that non-reproducibility is pretty much the default for most published papers.
I can't even reproduce CPU-Z benchmark results, on any hardware that I had, so... what's new?