Saturday, August 31, 2013

Let's Pay PhD Students More and Professors Less?

I just read a provocative post in The Guardian, titled "PhD: so what does it really stand for?"
The article focuses on the abysmal compensation of graduate students and recommends reducing salaries of professors in order to pay students more. Specifically, the proposal is to draconically cut the salaries of PIs:
A second option wouldn't hinder research, and might even enhance it: cut the salary of professors by half. If there are solid reasons for PhDs being paid half of what they deserve, then the same hold good for professors.)
What a terrible suggestion! Not terrible because students are paid enough  - what this article does not consider is that professors are already paid half.


In 2001, about to graduate from college, I turned down a programming position at a hedge fund.  Instead, I chose to do bioinformatics at Cold Spring Harbor Laboratory for a much lower salary.  I was excited about the possibilities of doing biological research using computational tools.  Two years later, I enthusiastically entered graduate school in molecular biology, with my salary dropping by half for the next six years. As a postdoctoral researcher at MIT, I am not even back to earning what I did ten years ago as a junior programmer with no skills or domain-specific knowledge.  In a commercial setting, my compensation would have kept pace with my knowledge and skills, but in academia, there seems to be a complete decoupling of the two. There are many sacrifices that academia requires, and I can spend pages moaning about them, but the suggestion to pay students more and professors less would probably exacerbate the problems and create a ton of new ones.

  1. In my opinion, one of the biggest concerns for research right now is the crisis in funding. (I have a a separate essay that I wrote about this a few months ago, and I will be posting it in 1-2 months for a good reason.) The funding crisis is so bad that we are about to lose an entire generation of brilliant professors. One of the causes of this is the glut of research labs - we need fewer labs. The Guardian article calls for more funding for PhDs to get more better students and to be able to fund more labs, and would make the crisis that much worse.
  2. If the prospect of being a professor already requires an enormous sacrifice and we are in danger of losing the best people from academia, how does telling aspiring professors that on top of all the problems, they will also be paid half for the rest of their lives help? Academia and professor positions are already barely competitive for the best and the brightest, cut life-long renumeration in half, and you have a fully unmitigated disaster.
  3. I realize that I have a terribly skewed perspective on the student body. Still, doing graduate work in Berkeley and a postdoc in MIT, if anything, I always felt inadequate compared to all of the amazing students. It never for a second seemed to me that we are not attracting top talent into PhD slots.
  4. I am far from convinced that graduate students are underpaid. 
I realize that #4 is an incendiary statement. So let me elaborate. Back as a student in Berkeley, I did an interesting comparison to answer exactly the question, "Are PhDs paid enough?" My wife had just graduated from a Physician Assistant program. The program was two and a half years in length, and after graduation, she instantly was able to start work with a salary of $70K. It was amazing - she started her school a year after me, finished basically half-way through my PhD, and started getting a salary that I wouldn't see for many more years. And she treats patients; her job has as many intrinsic rewards as my job. She is not sacrificing satisfaction at work for a higher paycheck.

So why do I think it's fair? Let's do some simple math:

Assume 6 years of PhD training with a stipend of $30K per year. At the end, my net plus is $180K. Meanwhile, for my wife, the 2.5 years of school have no stipend and cost a ton of money in tuition and books. At $30K per year of training, over the six-year period of my PhD, she is looking at -30*2.5+$70*3.5 = net plus of $170K. So over the 6-year period, I got $10,000 more than her in my "underpaid PhD". 

As a final comment, I have had several discussions with friends who argue that the way to fix the healthcare costs in United States is to cut the salaries of doctors. From a purely selfish position, that is bad for us because high cost of doctors increases demand for physician assistants. But the real argument against paying doctors less is the same as against paying professors less - the demanded sacrifice to get to the final position is so great, decrease drastically the final long-term compensation, and no one will go into medicine. (Through medical school, residency, fellowship - it takes you so long to get to be a practicing physician, and many start with $300-500K student loans hanging over their heads - that to also not compensate the doctors after the decade of training is just preposterous.)

Thursday, August 29, 2013

Do you have to publish in Cell/Nature/Science to get a professor job?

A decade ago, I asked Michael Eisen at the party celebrating the launch of PLOS Biology, "Mike, what if refusing to publish in Cell/Nature/Science hurts your prospects of getting a faculty job?"
Mike replied in his typical fashion, "Then we'll open a research institute for open access authors."

Four years later, I had been through too many conversations about the "sacrifice of submitting top work to PLOS Biology instead of CNS." It all boiled down to a classic Russian joke:

   A mental institute patient who thinks he is a grain is being discharged. Doctors are quizzing him, "Are you sure you are cured and you are not a grain?" The man says confidently, "A grain? Are you nuts? I am a person - I have hands, legs, a head. I am talking to you. Of course I am not a grain." So he is released, and 30 seconds later rushes back into the room slamming the door behind him in panic. "There is a chicken out there!!!!" The doctors are puzzled, "But you just told us that you are not a grain." The man replies, "Of course I am not a grain. I know that. But the chicken doesn't!"

And so it was with PLOS Biology. "I know my story is good. I know papers in PLOS Biology are not inferior to those in CNS. But the hiring committees don't believe it, and if I submit to PLOS, it will hurt my job prospects." Given this, I tried to measure whether submitting to PLOS really does undermine you in the faculty search.

As a proxy for postdoc->professor transition, I measured the frequency per journal of first authors subsequently becoming last authors in later years. I never got around to writing this up, and I think the analysis needs to be redone now with more open access journals and more authors. But it's still useful to share, and so I just uploaded the results to figshare.

I know all of the caveats. Those publishing in PLOS Biology in 2004 and 2005 could have been in labs where the PI is so well known and established that it didn't hurt them. More importantly, I did not ask where these people got the jobs, so it is possible that the PLOS Biology authors had fewer authors and ended up in less prestigious universities and had lower startup packages. Most importantly, this analysis says nothing about whether committing exclusively to open access journals and forgoing CNS entirely has a negative impact on downstream funding and tenure decisions.

I will write a separate blog post on the topic of open access commitment and who can afford it, but I do think that my analysis from 2007 shows something important. It shows that the postdocs from the Eisen lab who ended up in faculty positions are not the only people who can forgo CNS and still get a professor job.

(I would love to see someone repeat/expand this analysis. So, I am also posting all of the scripts and detailed documentation on figshare)