Postdoc Progress
This past weekend I spoke at a workshop at the FASEB meeting on how to manage postdocs in the lab. I go to a lot of these things, but this one seemed particularly promising.
People have complained for years (maybe even decades by now) about the quality of mentoring, or lack thereof, for postdocs and graduate students. Why is progress so slow? In part, it's because academia is an incredibly conservative institution - the incentives under which academics operate don't reward many kinds of process improvements, and there is very little turnover in the workforce, so a lot of times you are stuck with old ideas for decades until people retire. In part, though, the process of making institutional changes is complex.
What was encouraging about the FASEB meeting was that I saw signs that for the case of postdoc mentoring and management, the process is well underway:
Motivation: All too often one hears, "We have the best system of science education in the world -- why jeopardize that?" It's an excuse that is selectively applied, however. To fend it off, one needs a compelling reason. In the case of postdoc management, a lot of reasons have been put forth over the years. At this particular meeting, a reason that I think is particularly convincing was the common theme - that effective management of one's postdocs (and grad students) pays off in increased scientific output (disclosure: I am one of the sources of this idea). My friends in the real world are always amazed that there is anyone who doesn't see this as obvious, but my experience has been that it's a hard sell in some academic quarters. The National Academies and the Sloan Foundation both deserve a lot of credit for getting the ball rolling here.
Methods: Once an organization is convinced it needs to change, the next hurdle is figuring out how exactly to go about it. What is the best way to improve mentoring for postdocs and lab management in general? This next step involves a lot of documentation and training. A process needs to be laid out, tested, and revised / tuned / optimized. Training materials need to be developed, trainers need to be trained, and the training needs to be somehow incorporated into the non-curriculum for postdocs and their supervisors. The Burroughs Wellcome Fund and HHMI have really stepped up in this regard and have produced an outstanding lab management course. Maryrose Franko gave a nice presentation on the course.
Dissemination: All the motivation and methods in the world won't help if people don't know about them. Getting the word out to postdocs is pretty challenging since they are so widely diffused. Here the National Postdoc Association is playing an important role. They helped organize the workshop, and through their own networks are continually in touch with postdocs. Most of the time, these kinds of conferences involve postdocs and a few sympathetic senior people preaching to an audience of postdocs. At the FASEB meeting, it was mostly senior PIs talking to new PIs. A sympathetic audience, to be sure, but one that is a lot more likely to do something with the materials presented. Again, the Sloan Foundation has been instrumental in making this happen.
One particularly interesting sign that ideas about improved postdoc management are diffusing into the right places is in the bill reauthorizing the NSF, HR 1867:
SEC. 8. POSTDOCTORAL RESEARCH FELLOWS.
(a) Mentoring- The Director shall require that all grant applications that include funding to support postdoctoral researchers include a description of the mentoring activities that will be provided for such individuals, and shall ensure that this part of the application is evaluated under the Foundation's broader impacts merit review criterion. Mentoring activities may include career counseling, training in preparing grant applications, guidance on ways to improve teaching skills, and training in research ethics.
(b) Reports- The Director shall require that annual reports and the final report for research grants that include funding to support postdoctoral researchers include a description of the mentoring activities provided to such researchers.
I am guessing this addition, like HR 1453, has arisen from the input of a AAAS congressional fellow. I haven't looked at the Senate version of the reauthorization, so I don't know if there's anything similar there. However, the fact that mentoring requirements have made it this far is certainly encouraging.
-
on Fri, May 04, 03:05PM
The info about the bill reauthorizing the NSF as having aspects of postdoc mentoring and reporting is very exciting to me. Your analysis of why the academy (I'm in it) is so slow to change due to internal considerations is accurate (although there are also other factors including changing the attitude of the administration re: valuing faculty activities related to postdocs). Having external considerations coming from NSF and hopefully, NIH, will give a big boost to those within the academy who want to make changes in postdoc mentoring and management. At the recent NPA meeting in Berkeley, in a personal discussion with Peter Preusch, NIH Training Officier, he expressed the opinion that federal agency oversight of postdoctoral mentoring and management of postdocs supported on research grants was not likely to happen soon. It's exciting that what is currently being considered in this regard is contrary to his perspective!
-
on Mon, May 07, 12:05PM
We'll see what happens - I looked at the senate version of the NSF reauthorization and didn't see anything similar, so it could get cut out in committee. The NSF and NIH are fairly different, too -- the NIH doesn't have requirements for discussing broader impacts of work in grant proposals, do they? Regardless, the fact that the proposal made it this far is a good sign.
-
-
on Sun, May 06, 01:05AM
Hmmm, so the NSF now requires a paragraph of text in grant applications and annual reports where the PI claims to have provided mentoring? That's better than nothing, but I have to be skeptical about its having any real impact. There's an issue of culture that's really difficult to change, and while I wish I had something constructive to propose, I'm drawing a blank.
How about changing the standard acknowledgment of NIH funding in publications to require a precise description of the contribution of the grad student/postdoc? When PI's have one of those postdocs who is a pair of hands in the lab but doesn't write papers (the large majority, in my experience) let them say so. That has a slightly higher barrier to wishful thinking than does some boilerplate about "training in research ethics".
-
on Mon, May 07, 12:05PM
The requirement hasn't been enacted into law yet (and I'm not sure it will be). However, requiring a statement about mentorship in the proposal and in the annual reports does a couple of things: (1) it forces the PI to at least pay lip service to the idea of mentorship, and (2) it sets up at least the distant prospect of enforcement. I can't imagine that anything ever would be enforced, but still. The idea of defining what an undergraduate education should involve has universities all nervous about academic malpractice - I could see this kind of requirement creating the same kinds of concerns.
If this were to go through, I would guess that it would create a demand that the university level for postdoc offices to provide basic professional development services to cover institutions' backsides, and that would be a good thing.
-
-
on Tue, May 08, 05:05PM
It has been my experience that many Ph.D. advisors place little importance on management skill, and that most Ph.D. advisors are poor at management. If the NSF requires that applicants put in a broader interests paragraph on postdoc mentorship then some awardees will be granted permission to use postdocs to get their research accomplished while paying only lip service to educating prospective Ph.D.s instead of "ensur(ing) that research is fully integrated with education so that today's revolutionary work will also be training tomorrow's top scientists and engineers". There are ample examples of laboratories that are quite "top heavy" with more postdocs than students who are funded quite a bit through NSF, even though the emphasis has been on Ph.D. traineeship.
-
on Wed, May 09, 10:05AM
I propose more tangible changes that I think need to be made concerning postdocs. With Congress considering such issues, perhaps with some lobbying we can get a few changes in this year. First, the number of postdoc positions are increasing linearly with time, while the number of "staff" research position are either relatively flat, or declining. This needs to change.
Question: How many postdocs are there in the U.S.? My guess is about 75,000. Some references on these numbers would be useful. I know from this reference, "Enhancing the Postdoctoral Experience for Scientists and Engineers" they quote about 55,000 postdocs, but this data is now 10 years old (pre NIH budget doubling), and never included postdocs who obtained their PhDs outside the U.S., now a very large component of the postdoc population (I would estimate around ~25%). See:
http://books.nap.edu/html/postdoctoral_experience/
http://books.nap.edu/openbook.php?record_id=9831&page=5
Acadamia is not the only employer hiring postdocs: Non-profits, government, industry, research hospitals, etc., all have increasing percantages of postdocs on their staff. Take NIH for example. From what I am told, the Bethesday NIH campus has roughly 3,000 postdocs and 9,000 researchers. Can someone please provide more precise numbers here and the reference the source? So, if my numbers are correct, 30% of the researchers are postdocs at NIH! Why, because they are so inexpensiver relative to a staff government researcher and have little job security, so you can let them go with the end of a grant or contract. Such numbers are not uncommon. I would estimate that at many research labs, 20% to 25% of the PhD-level research staff are now postdocs. Most labs have attrition rates of maybe 2-4% per year of research staff (not postdocs) so labs have anywhere from 5 to 10 the number of postdocs needed to compensate for attrition. To the extent that other employers have similar ratios, I imagine it is less in industry, this makes for the bottleneck for transitions to staff positions. For universities, the situation is obviously much worse. I know of one theory professor in physics who has NINE postdocs!
Universities could hire more "senior scientists", and many do. But as this article points out in the field of particle physics, this is an expensive proposition in a time of perpetually tight budgets:
http://mysite.verizon.net/vze2sk2i/spn/ref07/senscientists.pdf
I believe that Congress needs to audit the number of postdocs, and staff-to-postdoc ratios, in the government labs, and get the numbers down to around 10%. Converting postdocs to staff, will cost money, but it is certainly more tangible and better long-term solution than lobbying for better "mentoring". Faculty to postdoc ratios should also be examined.
A prelude to the next topic is postdoc stipends/salaries/benefits, they are far too low, and are one of the main reasons postdoc numbers are increasing with time. You can hire about 3 postdocs for every staff researcher, when you include overhead, FICA, and 401K benefits. This needs to change. Again, Congress has asked the NAS to keep track of this for biomedical researchers, but even the NAS committees think supply of PhDs in the biomedical field is in balance with demand.
See: http://grants.nih.gov/training/nas_report/index.htm
Bob
-
on Wed, May 09, 12:05PM
Hi Bob--
There are a couple of ways one can estimate the total number of postdocs. First, you can look at the Survey of Doctorate Recipients, a biannual survey of about 7% of the people who earned their PhDs in the US. That will give you a pretty big undercount because, as you say, it misses all the postdocs who earned their PhDs outside the US (about half of all postdocs). A second method is to look at the NSF's annual survey of graduate students and postdoctorates. This is a department by department head count of postdocs and gets both people who earned their PhDs in the US and elsewhere. However, this second survey just counts people at academic institutions and misses people who have postdocs at government labs and at nonprofit biomedical research centers.
From WebCASPAR, I find that as of 2005, there were 48,653 postdocs at academic institutions. Sigma Xi surveyed the NIH postdocs in our survey in 2005; there were about 3,000 all told. There are probably a couple of thousand more spread out over the various national labs, and a few thousand at places like Scripps, Dana Farber, and so on. So maybe 55K-60K all told.
The NIH doubling created about 4,000 new postdocs -- see here
Having Congress dictate ratios of postdocs to staff scientists sounds to me like a recipe for lots of trouble. Imagine if, say, lab automation changed staffing needs dramatically - how long would it take for Congress to catch up? The H-1B numbers are random and capricious because they have to be manually adjusted by a slow and painful Congressional intervention. Imagine if that kind of time lag were involved in basic staffing or salaries! I was on an NSF postdoc in the early 90's; the amount was never adjusted for inflation, so it had been the same for a decade or more by the time I had gotten one. That's what you'd get if you put salary decisions in the hands of Congress.
I think there are things Congress could usefully do that are less heavy-handed than having them intervene directly in staffing and salary decisions. Think about kinds of things that would make labs want to do the right thing. Transparency measures would help a lot.
Perhaps some liability-like things as well. I'm sort of surprised that there have been so few high-profile cases of postdocs suing their PIs - after all, a federal circuit court has ruled that PIs have a fiduciary duty to their postdocs.
Also, Congress is rarely a place where innovation takes place. At best, they will pick up successful innovations from elsewhere. Professional Science Master's programs were created and funded by Sloan, and now the senate is allocating money to have the NSF pick up the tab. What kinds of similar practices might lend themselves to such extensions?
Geoff
-
-
on Thu, May 10, 07:05AM
Geoff,
From your comments below, you seem to exhibit a lack of appreciation for the cause and effect concerning policy matters on Capitol Hill. As you must know, Congress responds to its constituents (usually only large numbers of well organized constituents), and/or powerful lobbyists. Right now, young scientists are not well organized politically (despite the growing National Postdoc Association) nor are they powerful. With a lot of effort, this could perhaps change, but it remains to be seen. Concerning your response to my suggestions:
Geoff Davis on Wed, May 09, 12:05PM wrote: "Having Congress dictate ratios of postdocs to staff scientists sounds to me like a recipe for lots of trouble. .... The H-1B numbers are random and capricious because they have to be manually adjusted by a slow and painful Congressional intervention. Imagine if that kind of time lag were involved in basic staffing or salaries!"
The "H1b numbers" are not "random and capricious" if you followed the lobbying efforts of high-tech lobbyists and university presidents (the AAU) in the late 1980s, and again in the late 1990s, which were countered less effectively by concerned scientists and engineers (S&Es). The H1b quota remains set at 65,000 today precisely because displaced and concerned S&Es have lobbied against further increases, so now Congress knows this is a controversial, two-sided, issue. There have been several attempts by Congress to slip H1b increases in un-related appropriations bills last year and this year, which have failed because other Congressman and Senators sympathetic to S&E labor have stepped in. So, writing letters to Congress in large numbers can have an impact.
"I was on an NSF postdoc in the early 90's; the amount was never adjusted for inflation, so it had been the same for a decade or more by the time I had gotten one. That's what you'd get if you put salary decisions in the hands of Congress."
Again, the stipend level for NSF postdoc fellowships are in the hands of the NSF for the most part, NOT Congress. Yes, one could try to lobby the NSF to increase the stipend levels if postdocs were in a position to be so politically inclined. If that fails, I would suggest trying to lobby your local Congressman and Senator to see if a letter from him to the NSF Director has an effect. It may take hundreds of postdocs writing similar letters to get anything going. In the end, maybe nothing will happen, but at least "we" tried. The National Postdoc Association (NPA) is trying to put some light pressure on the NSF and NIH concerning postdoc stipends, but has been more focused on increasing membership and applying pressure at the local level by increasing the number of local postdoc associations. The leadership of the NPA was reluctant to "lobby" Congress more directly due to its non-profit status.
"I think there are things Congress could usefully do that are less heavy-handed than having them intervene directly in staffing and salary decisions. ......Perhaps some liability-like things as well. I'm sort of surprised that there have been so few high-profile cases of postdocs suing their PIs - after all, a federal circuit court has ruled that PIs have a fiduciary duty to their postdocs."
I am not sure what could be more "heavy-handed" than having postdocs suing their PIs to affect change? I was thinking a well-written one-page letter to their Congressman might be more appropriate.
Well, I have to run to local conference. This is an important issue. I hope we can focus on it. I have more info concerning postdoc stipends levels that I wanted to pass on, but I ended up spending time rebutting your input.
I leave you with this link. Does anyone know what the status of the NSF proposed postdoc stipend increase is?
What is their level today?
More to follow....
Bob
-
on Thu, May 10, 12:05PM
Bob,
I think you misunderstand me RE approaches to the situation. I think it is more effective to look to government for tools for solving problems rather than for solutions. That's a longer discussion, though - I'll try to get to it when I return from Portland.
As for postdocs lobbying, try to put yourself in the shoes of a Congressman:
- Postdoc salaries, about $35K in 2004, are in line with the average US wage
- Postdocs are temporary training positions that (ostensibly) prepare people for much higher salaries in industry or, in some cases, as faculty members.
- The more postdocs cost, the more it costs to pay for health care research.
Sure, $35K/year sucks for someone who has spent all those years earning a PhD. But if there is a real training component, maybe it's not so bad. The way I see it, in most cases the money isn't the problem - it's the relative lack of training.
As for constituents, I'd guess that there are a lot more voters who want to keep health care costs down (me included!) than there are postdocs who want to get paid more.
Geoff
-
on Thu, May 10, 12:05PM
Oh yes, RE stipend amounts, here's what I dug up on my old NSF postdoc:
http://www.nsf.gov/pubs/2005/nsf05510/nsf05510.htm
$4K/month is about twice what I got, and it's considerably more than Rita Colwell's proposed $40K minimum.
The funding agencies pay decent stipends when they fund people directly (now that they have gotten the health insurance situation mostly straightened out).
-
