Peter and I have just returned from "Evolving Demands in Graduate Education, Training and Career Development for Future STEM Professionals," a workshop hosted by the White House Office of Science and Technology Policy.
The meeting was in some ways a testament to the potential of young scientists and engineers to influence the community. Sharon Hays, the organizer, is a life scientist who earned her PhD during the last big downturn for scientists in the mid 1990's. As a result of her experiences in grad school and on the market during the 90's, she has had a longstanding interest in science careers, and she is now in a position to do something about them: she is the Deputy Director of OSTP.
Around the table were most of the key people involved in the support and training of graduate students in the United States along with a number of thought leaders from academia and industry. If ever there was an opportunity to explore this issue with all the stakeholders present – this was the time.
The meeting was also a reminder of the extreme inertia of the scientific world: 10 years ago, many of the same people (Sharon, Peter, Finley Austin, and myself) assembled at a similar meeting hosted by Congressman and physicist Vern Ehlers. As Sharon noted in her closing remarks last week, many of the slides shown at the meeting in 2007 could have been taken straight from the 1997 meeting.
Sharon opened the meeting with several questions:
- Why are graduate students trained with a near singular focus on becoming academics when smaller and smaller numbers of these students actually go on to being academics?
- Why does it take so long to earn a PhD? MD/PhD programs prove that the time can be made shorter: in 7 years, everyone earns both a PhD and an MD.
- Can we better predict supply and demand for scientists and engineers?
- How can we fix problems without undermining the science and engineering enterprise that is dependent upon low-cost graduate students and postdocs?
A few quick highlights of the meeting:
Are producing too many scientists? People are still asking this question, and there was considerable disagreement on the answer.
- Michael Teitelbaum made a compelling empirical argument for an oversupply - young scientists' wages are low and many people have considerable difficulties finding jobs - and raised concerns about the upcoming NSF budget doubling.
- Charlotte Kuh presented an interesting argument that overproduction was built in to the system as a result of incentives for the various actors in the system being incompatible with desired outcomes.
- All the industry people (Pfizer, Dow, Hoffman-La Roche) reported considerable difficulty in finding suitable hires.
- David Skorton, President of Cornell, argued vigorously that the problem is simply that not enough Federal money is being invested in science and engineering.
How are things in the trenches? We had a great panel of grad students and postdocs who related their own experiences. It's not clear how representative the 4 people were of grad students on the whole, but to the extent that they are, some positive things are going on. Some common themes that emerged from their stories:
- Most of the panelists had not gone straight through from undergrad to graduate studies and had pursued interesting paths in between. They went to grad school thinking that a PhD would lead to enjoyable work and give them control over their careers.
- All reported having received little to no career guidance along the way.
- Most were planning to pursue non-academic careers because of perceptions that life as a professor is undesirable - it's a career that leaves no time for family or anything else but work.
- Despite sounding happy about choosing not to pursue academic careers, most sounded a little defensive about their nonacademic goals.
- 2 of the 4 grad students were strongly influenced by NSF’s Research Experience for Undergraduates (REU) program. The REU gave them a window on research science that they found really enticing. Several wished they had had the opportunity to have explored something similar in industry.
- Several reported receiving some positive structured career guidance in the form of a disciplinary society workshop. All wished they had had such an experience earlier in their graduate careers and thought that it would have been most effective earlier on.
- All wished they had had more exposure to opportunities in industry early on in their educations.
My take away was that there are a few real opportunities for OSTP to improve things:
1) The recent NIH budget doubling is a useful negative case study. Some Federal resources invested in understanding what went wrong and in better resource planning could go a long way toward preventing similar problems with NSF's upcoming doubling and with future human resources planning in general. Funding an external agency like the Bureau of Labor Statistics to do the work would be particularly useful: BLS has relevant expertise that is simply not a core competency of NIH/NSF/DOE/NIST/etc; and the BLS budget would not be affected by their findings about national needs for S&E, so the conflicts of interest that affect NIH/NSF's existing resource planning (projections of future work force needs influence Congressional appropriations to NSF/NIH/etc) would go away.
2) There was broad consensus at the meeting that giving students more exposure to careers in industry would be very productive. One hurdle is that it is expensive and time consuming for universities to establish and maintain the kind of industrial ties that the envisioned internships / interactions would require. Having NSF and especially NIH (since NSF already does this kind of thing to a limited extent) allocate some substantial resources in this direction would be another way OSTP could have a big impact.
3) It would be useful to fund structured guidance (i.e. workshops / classes rather than occasional suggestions from an advisor). Disciplinary societies are filling some of the vacuum left by NSF's decision to stop analyzing its own data on S&E's, but the quality and timeliness is all over the map. Some funding to help disciplinary societies provide good data to students and faculty members would be useful as would funding for institution-level services at graduate student and postdoc offices.
Two more ideas from Peter:
4) Funding graduate students directly on research grants (as opposed to supporting them on fellowships) was discussed at several points. Directly coupling PhD production to research funding is nice for PIs because it gives them a ready and eager source of cheap labor and total control of that labor. However the direct coupling ties PhD production to R&D funding levels – which invariably creates a boom/bust cycle of PhD production. Michael Teitelbaum made a point of this with respect to NSF: for every one graduate student supported on an NSF fellowship, 12 students were supported directly on grants. If you double NSF funding over the next 10 years you are signing up for a substantial increase in PhD production. Teitelbaum argued that this NSF should lower this ratio substantially (presumably by funding more fellowships relative to grad student slots on R&D proposals). This idea of decoupling the funding for graduate students from the funding of research is not new – it was a central idea that came out of the LAST crisis. Back then, the idea was to empower graduate students to make their own choices as to who to train with and what projects to undertake. But if you look at the ratios of fellowships to support through research funds since 1990 the proportion on fellowship support hasn’t budged.
5) A mismatch between supply and demand. Despite the abundant data showing the indicators of overproduction of PhDs in some fields (this time – Life Sciences) the folks from industry insisted that attracting skilled scientists and engineers was a challenge. Several industry folks explained that it was not simply a question of supply of PhDs but rather PhDs with the suitable skills and INCLINATION to consider industry jobs. One of the industry reps highlighted the “Rodney Dangerfield” syndrome that many technical managers feel: they can’t get the respect of their peers in academia who train and counsel graduate students. Furthermore, the supply/demand issue is very field-specific. Presently in the Life Sciences we are producing too many academic-oriented PhDs: WAY too many for the available funding streams. However, in Computer Sciences, PhD grads (at least according to the Dean of one top-tier department) are beating away the recruiters with sticks.
One thing that has changed since the last PhD glut is the avenues for communication. Peter and I have invited the Roundtable participants to join this blog to share their observations and ideas. We will also post the presentations that are made available.
