4 ms·

The cause of the problems in academia is hinted at near the end: > the truth is I understand why they have to do things that way, if they don't get published,

The cause of the problems in academia is hinted at near the end: > the truth is I understand why they have to do things that way, if they don't get published, they don't get money to continue work that should theoretically be beneficial down the road! Scientists optimizing for what is likely to be published (read: what is likely to earn money), sacrificing elsewhere. Maybe if we payed them enough so that they don't have to do this to survive, then we wouldn't have this issue.

munin 8y ago
> Scientists optimizing for what is likely to be published (read: what is likely to earn money), sacrificing elsewhere. Maybe if we payed them enough so that they don't have to do this to survive, then we wouldn't have this issue. How do you do this in a society with scarce resources? With scarcity, you need to have a measure of success, otherwise everyone could just say they were a scientist and get a six figure salary to do nothing. Any metric you choose will be gamed. This worked hundreds of years ago because the only people that could get educated were the wealthy, so we didn't need to worry about supporting scientists with public funds, they supported themselves.
- tonysdg 8y ago
  > With scarcity, you need to have a measure of success, otherwise everyone could just say they were a scientist and get a six figure salary to do nothing. Any metric you choose will be gamed. As a current grad studrnt, I'm calling BS. We (meaning academics, even the ones gaming the system) _know_ what strong, reproducible, well-designed research projects look like. It doesn't take a genius to develop an experiment, run it, document what happened, and publish _all_ of the results. Hell, you learned the basics in middle school -- pick an independent variable, tweak it, then see what happens to the dependent variables. Now it's obviously not that simple -- isolating just one independent variable is practically impossible much of the time -- but the concept of documenting EVERYTHING that you tried should be obvious. There's absolutely zero impetus to share all of that information, however, especially when you only have 12 pages in a conference paper to write up everything. So instead you go fishing, and you run a boatload of tests, and you see which ones "solve" the problem, and you write up just those tests, ignoring all the failures and the false starts. It pisses me off, to be honest, and it's just one of the many reasons I'm trying to get out of here as soon as I can.
  - munin 8y ago
    I'm also a grad student. We do know what good research looks like, that's not the problem, at least not directly. We're both told "go do good research" and then we go and each write one paper. The government then says "okay this is fine, but next year, I can only fund one of you, so, now whoever brings me the most papers next year will keep their job, and the other will hit the bricks." This is where the problem starts.
- Fomite 8y ago
  It has been suggested that a lottery system for grant proposals would actually both work better and be more fair than our current system.

munin 8y ago

> Scientists optimizing for what is likely to be published (read: what is likely to earn money), sacrificing elsewhere. Maybe if we payed them enough so that they don't have to do this to survive, then we wouldn't have this issue. How do you do this in a society with scarce resources? With scarcity, you need to have a measure of success, otherwise everyone could just say they were a scientist and get a six figure salary to do nothing. Any metric you choose will be gamed. This worked hundreds of years ago because the only people that could get educated were the wealthy, so we didn't need to worry about supporting scientists with public funds, they supported themselves.

tonysdg 8y ago
> With scarcity, you need to have a measure of success, otherwise everyone could just say they were a scientist and get a six figure salary to do nothing. Any metric you choose will be gamed. As a current grad studrnt, I'm calling BS. We (meaning academics, even the ones gaming the system) _know_ what strong, reproducible, well-designed research projects look like. It doesn't take a genius to develop an experiment, run it, document what happened, and publish _all_ of the results. Hell, you learned the basics in middle school -- pick an independent variable, tweak it, then see what happens to the dependent variables. Now it's obviously not that simple -- isolating just one independent variable is practically impossible much of the time -- but the concept of documenting EVERYTHING that you tried should be obvious. There's absolutely zero impetus to share all of that information, however, especially when you only have 12 pages in a conference paper to write up everything. So instead you go fishing, and you run a boatload of tests, and you see which ones "solve" the problem, and you write up just those tests, ignoring all the failures and the false starts. It pisses me off, to be honest, and it's just one of the many reasons I'm trying to get out of here as soon as I can.
- munin 8y ago
  I'm also a grad student. We do know what good research looks like, that's not the problem, at least not directly. We're both told "go do good research" and then we go and each write one paper. The government then says "okay this is fine, but next year, I can only fund one of you, so, now whoever brings me the most papers next year will keep their job, and the other will hit the bricks." This is where the problem starts.
Fomite 8y ago
It has been suggested that a lottery system for grant proposals would actually both work better and be more fair than our current system.