3 ms·

Given what I learned in recent years about the scientific process I'm always a bit wary of making this a "spend XXX money to get YYY results". We have issues l

by hannob 8y ago

Given what I learned in recent years about the scientific process I'm always a bit wary of making this a "spend XXX money to get YYY results". We have issues like that still a sizable amount of medical studies are unpublished because they don't produce the desired result the funder wants, that experts think a vast amount of studies are simply research waste due to poor methodology or that the majority of pre-clinical cancer studies can't be replicated. How much better could we be at tackling diseases if science wouldn't have these problems? Sure, more funding is one part of the equation. But I increasingly get the feeling that if modern science wants to make a significant step forward they need to clean up their methodological problems. And that will not necessarily cost more, in many situations it may even cost less. (E.g. don't pay for studies that won't get published later or that are methodologically so weak that they won't produce actionable results.)

FabHK 8y ago
> don't pay for studies that won't get published later or that are methodologically so weak that they won't produce actionable results. Most typically, to make studies methodologically stronger, you require a bigger sample, ie more subjects, to increase power while maintaining or even improving significance. And that would make it more expensive.
- leoc 8y ago
  But it's uncertain whether simply pouring more money into the funding hopper would generally result in studies with larger sample sizes, or whether it would just result in a larger number of surveys with a sample size similar to the deficient ones we see today, as a larger number of scientists manage to cling by their fingertips to the gravy train. After all, there's presumably little good reason why we couldn't have fewer but better studies today, at existing funding levels.
  - Broken_Hippo 8y ago
    You could be reasonable and make larger sample sizes and so on requirements. While it isn't certain that simply pouring money into funding hoppers results in more studies, it is pretty certain that larger sample sizes (and things like this) cost more money. It is prudent to give enough for these things. If you want to have more accountability with money, build it in. You might find it costs more or less than current funding levels - I'm going to guess costs will vary with sort of funding. Use equipment funds for equipment. Give researchers and their staff actual, decent, steady wages not reliant on results. (A "negative" result isn't negative, after all, for it disproves something). And so on.
- Brakenshire 8y ago
  I spoke to a medical researcher who had worked both in public and private labs during their career, they said when they worked in public academic labs there was incompetence to the extent that no-one understood how to properly calibrate certain lab equipment. Working in private research, they said public research generally could not be replicated, and unless from a known lab was considered borderline useless. I think it goes way beyond what you’re talking about. In large sections of academia there’s a very Darwinian system where researchers are fighting for their careers and labs are trying to keep their heads above the water, and everyone needs sexy results now. I don’t think such a system can lead to robust, unbiased research by default.
antt 8y ago
There needs to be a journal for negative results. >We tried X it didn't work. Would be invaluable in a lot of fields where knowing what's been tried is more important than knowing what hasn't.

by hannob 8y ago

FabHK 8y ago

> don't pay for studies that won't get published later or that are methodologically so weak that they won't produce actionable results. Most typically, to make studies methodologically stronger, you require a bigger sample, ie more subjects, to increase power while maintaining or even improving significance. And that would make it more expensive.

leoc 8y ago
But it's uncertain whether simply pouring more money into the funding hopper would generally result in studies with larger sample sizes, or whether it would just result in a larger number of surveys with a sample size similar to the deficient ones we see today, as a larger number of scientists manage to cling by their fingertips to the gravy train. After all, there's presumably little good reason why we couldn't have fewer but better studies today, at existing funding levels.
- Broken_Hippo 8y ago
  You could be reasonable and make larger sample sizes and so on requirements. While it isn't certain that simply pouring money into funding hoppers results in more studies, it is pretty certain that larger sample sizes (and things like this) cost more money. It is prudent to give enough for these things. If you want to have more accountability with money, build it in. You might find it costs more or less than current funding levels - I'm going to guess costs will vary with sort of funding. Use equipment funds for equipment. Give researchers and their staff actual, decent, steady wages not reliant on results. (A "negative" result isn't negative, after all, for it disproves something). And so on.
Brakenshire 8y ago
I spoke to a medical researcher who had worked both in public and private labs during their career, they said when they worked in public academic labs there was incompetence to the extent that no-one understood how to properly calibrate certain lab equipment. Working in private research, they said public research generally could not be replicated, and unless from a known lab was considered borderline useless. I think it goes way beyond what you’re talking about. In large sections of academia there’s a very Darwinian system where researchers are fighting for their careers and labs are trying to keep their heads above the water, and everyone needs sexy results now. I don’t think such a system can lead to robust, unbiased research by default.

antt 8y ago

There needs to be a journal for negative results. >We tried X it didn't work. Would be invaluable in a lot of fields where knowing what's been tried is more important than knowing what hasn't.