A modest proposal for fixing the scientific system

[debateg]

---

Members do not see this ad.

There has been a flurry of discussion about the problems MD/PhD's face as they progress in their careers. Some seem frustrated that few have proposed solutions. A recent white paper from the NIH advocated changing funding for trainees from investigator grants to training grants: http://www.nationalpostdoc.org/inde...news/669-2012-nih-biomedical-workforce-report. The end result would be that training grants would be administered by departments and divisions and independently reviewed based on how successful their trainees are in terms of publication, time of training, and career outcomes. Currently investigator grants are reviewed by how successful their science is, not how their trainees do. The end result is that trainees are generally not treated well, and there is no accountability for churning out PhDs who find their career options quite limited.

Do you think making this change would address some of the issues? What problems do you see with this?

It's worth discussion because this is a change that actually could be implemented, and you could contact people at the NIH to advocate for it.

 

[solitude]

There has been a flurry of discussion about the problems MD/PhD's face as they progress in their careers. Some seem frustrated that few have proposed solutions. A recent white paper from the NIH advocated changing funding for trainees from investigator grants to training grants: http://www.nationalpostdoc.org/inde...news/669-2012-nih-biomedical-workforce-report. The end result would be that training grants would be administered by departments and divisions and independently reviewed based on how successful their trainees are in terms of publication, time of training, and career outcomes. Currently investigator grants are reviewed by how successful their science is, not how their trainees do. The end result is that trainees are generally not treated well, and there is no accountability for churning out PhDs who find their career options quite limited.

Do you think making this change would address some of the issues? What problems do you see with this?

It's worth discussion because this is a change that actually could be implemented, and you could contact people at the NIH to advocate for it.

I think this is a good idea. I believe another suggestion in that white paper was to set a hard cap on the number of years that would be funded for PhD. After those years were expired, either the student would graduate, or the PI would have to divert money from their main research grants to pay for the trainee's funding. To make this more potent, they should mandate graduation.

IMO there is an over supply of PhDs. I would cut funding for PhD training. Supply and demand should restore some balance in the academic and industry marketplace.

A corollary is to make quals much more relevant. Nowadays, virtually everyone passes their quals. I think that they should postpone quals until about 3 years in. At that point, with only 2 years left on your training grant, it should be pretty clear whether somebody is going to complete his PhD. Let the quals sort out who deserves to graduate and who doesn't, and award a Masters to those who don't make the cut. They could even introduce two rounds of quals, one after year 1 and another after year 3, with attrition after year 1 due to people who simply can't hack it, and attrition after year 3 due to people who won't finish by 5 years.

I believe these would be positive steps, but that the scientific enterprise is more generally dysfunctional. Most papers are erroneous and not reproducible (e.g. see 2012 Nature paper where a team of 100 from Amgen could only reproduce 6 of 53 papers published in prestigious journals). I think that the NIH should form an Institute for Replication, and a huge chunk of the budget should go towards replication studies. They could either say that you can't publish until we've reproduced your paper, or they could say: go ahead and publish it, but if it proves not to be reproducible, you have to pay back the money and retract the paper. We need to move away from a system in which all that matters is publishing fashionable papers in trendy journals, and towards a system in which the quality of a career rests on the validity of one's findings.

Another easy step would be to discourage mega-papers. It just takes too damn long to publish nowadays, especially "high-impact" papers. I think it slows down science, and places too much emphasis on forming a coherent story that is mostly data massage.

Similarly, incentivize publication of negative data and replication studies. Easy to do: just consider them of value during grant applications.

Ensure reproducibility by mandating that all NIH-funded data be made publicly available on a weekly or monthly basis. People couldn't get by with data obfuscation if all primary data was easily referenced online.

I have more off the wall ideas. Probability of acceptance: zero. The sad fact of the matter is that the people in charge are happy with the current system. Where is their incentive to change it? You will notice that none of the "recommendations" from that panel - which met for 2 years - have any teeth at all. It was really just a show trial to ease discontent. The impetus will have to come externally, i.e. the public and especially Congress.

 

[dl2dp2]

There has been a flurry of discussion about the problems MD/PhD's face as they progress in their careers. Some seem frustrated that few have proposed solutions. A recent white paper from the NIH advocated changing funding for trainees from investigator grants to training grants...

I actually think a much easier fix is to just give smaller chunks and have more frequent reviews. While this does add work for everyone, at least things wouldn't be as desperate and random. I.e. funding 50% of the research in a given cycle, but only give 50k to each grant...something like that...Although I think that idea is starting to get implemented...

 

[Jorje286]

Slash those PhD positions. This looks very simplistic but I'm not sure why it would be hard to implement. The PhD gives false hope for many people (in terms of prestige certainly), and it will be too late for many when they realize that the degree in itself is not that worthy. The number of PhD positions should basically be commensurate with the number of potentially funded faculty positions. In that scenario students may have to work more time as research assistants to prove themselves before they get enrolled in a PhD position, and publications will probably be a must.

 

[Neuronix]

Slash those PhD positions.

I agree with you Jorje. The amount of funding for graduate students should decrease, and that should be re-allocated to post-docs and junior faculty.

The practical problem is: who will be the low paid labor that supports the academic research enterprise? The reality of funding is that there is so much allocated for labor in the grant, and that goes to pay for the labor. Career technicians command a much higher salary. PIs don't usually do their own experiments. PhD students and post-docs get paid far less with the goal of moving on to a higher paid position. If you take away the students (either in a program or wanting to get into a program), the profzi scheme (http://www.phdcomics.com/comics/archive/phd030909s.gif) will collapse.

 

[debateg]

The practical problem is: who will be the low paid labor that supports the academic research enterprise?

Let's say that grad students and postdocs were supported only by training grants, not individual RO1's and the like. That could be accomplished by shifting money around. Now let's say that some of those training programs do not give good outcomes for their trainees. Upon peer review, those training programs get their grants cut or discontinued altogether. Now University XYZ suddenly has no postdocs and no grad students. The only ones left doing work are clinical fellows, technicians, and staff scientists.

Clinical fellows suddenly become much more valuable. There is intense competition to attract them. Technicians also become much more valuable. But most importantly, staff scientists become the work-horses of the lab. Labs that have sufficient grants will hire them, and all those unemployed PhDs now have a bunch of new jobs created. That comes at a cost of less money towards reagents, etc.

The end result is that science suddenly became more expensive to perform. Like any other industry, if you take care of your people, it costs money. Science is no different.