3 ms·

>One example of the principle is this: If you’ve made up your mind to test a theory, or you want to explain some idea, you should always decide to publish it wh

by spats1990 7y ago

>One example of the principle is this: If you’ve made up your mind to test a theory, or you want to explain some idea, you should always decide to publish it whichever way it comes out. If we only publish results of a certain kind, we can make the argument look good. We must publish both kinds of result. Isn't non-publication of studies/research (for various reasons) a massive problem in a lot of different disciplines now? [1], for example. [1]https://pediatrics.aappublications.org/content/138/3/e20160223 https://pediatrics.aappublications.org/content/138/3/e201602...

bonoboTP 7y ago
Mainly because reviewers won't accept uninteresting results. In many fields it's hard to differentiate a negative result from just a clumsy researcher. In computer science a negative result usually generates a flood of questions like "okay, but why would you do it like that, couldn't you try this or that etc." and the conclusion is that the researcher just didn't manage to get it to work because they didn't use the right techniques. Essentially you'd have to argue that the method you tried should work, based on what we already know, but surprisingly it doesn't work. This type of argument is very hard to carry through. The opposite is much more publication friendly: "nobody would have thought that this could ever work like this, but we now show that this novel idea does work very well". That's the type of thing for award-winning research.
- lallysingh 7y ago
  Publication -- somewhere -- of interesting negative results would be invaluable specifically for the discussion of how the experiment went wrong. Experiment design is tricky and it's not something you get a lot of exposure to except through the experience of others you interact with.
  - sifar 7y ago
    Eliezer Yudkowsky talks about this in Inadequate Equiibria [1] when he analyzes academic research. short answer - here you need researchers and grant makers to simultaneously decide to do what is not in the best interest for them. [1] https://equilibriabook.com/an-equilibrium-of-no-free-energy/ https://equilibriabook.com/an-equilibrium-of-no-free-energy/

bonoboTP 7y ago

Mainly because reviewers won't accept uninteresting results. In many fields it's hard to differentiate a negative result from just a clumsy researcher. In computer science a negative result usually generates a flood of questions like "okay, but why would you do it like that, couldn't you try this or that etc." and the conclusion is that the researcher just didn't manage to get it to work because they didn't use the right techniques. Essentially you'd have to argue that the method you tried should work, based on what we already know, but surprisingly it doesn't work. This type of argument is very hard to carry through. The opposite is much more publication friendly: "nobody would have thought that this could ever work like this, but we now show that this novel idea does work very well". That's the type of thing for award-winning research.

lallysingh 7y ago
Publication -- somewhere -- of interesting negative results would be invaluable specifically for the discussion of how the experiment went wrong. Experiment design is tricky and it's not something you get a lot of exposure to except through the experience of others you interact with.
- sifar 7y ago
  Eliezer Yudkowsky talks about this in Inadequate Equiibria [1] when he analyzes academic research. short answer - here you need researchers and grant makers to simultaneously decide to do what is not in the best interest for them. [1] https://equilibriabook.com/an-equilibrium-of-no-free-energy/ https://equilibriabook.com/an-equilibrium-of-no-free-energy/