|
|
|
|
|
|
by kurthr
972 days ago
|
|
|
I'm sure you'll find lots of people at her school who know the game, but for everyone else who hasn't heard. In order to get good grants you need to run "successful" experiments, which means you need to propose to find something interesting, but that means you want to already know there's something interesting there. That means you really need to bootstrap to have already run enough of the experiment that you know where it's going to go, before you write the grant for the money to run the experiment (and analyze the results, write papers, pay grad students, etc). You also need the money to run enough of the experiments for the new proposals you're going to write next so that they are "successful". Running experiments and not getting interesting results will lead to not getting future grants, because those handing out money have memories. Second thing. The best way to get tenure is to have such a huge group working for you (e.g. 30 grad students), because the provost will be terrified there isn't room for other grant writers in the department to get the money to support them and the lab you've built. Of course managing this is such a full time job before even teaching, that you'll be totally overwhelmed and delegating almost all of the research, papers, and grant writing to senior students. Once you get tenure you can slim back down to about 5 students. Also, if you come from a national lab or somewhere that gives you experience writing proposals (and preferably reviewing them) that is a huge advantage since you know exactly what the other reviewers are looking for. Once you write successful proposals and papers, you'll be invited to the committees and review other papers for important journals in your field. Of course a lot of this is again a bootstrap and treadmill problem. |
|
|
There's another factor too. More students means more papers and your h-index goes up. Were we to suppose that all students were equal and that there is a noise associated with the likelihood of publication and another noise with the number of citations then if either of those noise variables are large, you should maximize quantity over quality. Because you're simply increasing the odds that you'll hit a jackpot.
In fact, if you pull the data from csrankings.org you'll find that the first 30 (all I pulled because I'm lazy and not a web person) school's rank is practically a function of the number of publishing professors at that university. So the "more workers = more better" tactic actually scales from lab to department. If we look a little harder, I think we can all see the limitation of the metrics being used here and why they're so easy to hack. More importantly, why these metrics result in a dominating momentum force (aka. rich get richer). Maybe we should start reevaluating how we are evaluating systems. After all, it is neither fair nor an efficient usage of resources. If we're going to continue the trend though, the only solution is to allocate more resources... (which to be fair the pool of available resources is increasing year over year, but the allocation isn't)