Saturday, March 20, 2010

Are 'User Facilities' Killing Science?

The User Facility is, in my field, absolutely necessary. But this wasn't always the case.

Once upon a time, in the "golden days" of nuclear physics research, every large - no, I mean every research active - university had their own small accelerator. Hundreds and thousands of papers were published in the 60's and 70's (even into the 80's) by academics working with small cyclotrons or electrostatic accelerators in their department basements, with reasonably cheap detectors, simple electronics, and data acquisition "systems" that ranged from computers to mere photographic plates. It was easy to become familiar with your own particular setup, which in turn meant you produced better results because you knew the experimental strengths, weaknesses and limits. Sure, we were limited in energy, but at the time it didn't matter: the field was still new, still fresh and full of endless possibilities. You'd measure a reaction, not because it was important in Wolf-Rayet stars or because it populated the highest spin state, but because no one had done so before. Everything was novel.
At some point, however, the possibilities began to be exhausted. There was only so much you could do with light ion beams at low energy on stable targets. The field needed higher energies in order to progress, and that cost money.
As running the little basement accelerators slowly produced fewer and fewer publications, more and more universities and other small facilities shut them down. They became unnecessary expenditures, line items on university budgets that were no longer allocated. Only a portion of the existing facilities made the decision to put in the capital to expand their capabilities. And when this happened, nuclear physicists from the smaller places, whose facilities had been shut down, started traveling to the larger facilities.
The trend propagated with the need for even higher energies and new beam species. The advent of the radioactive ion beam (RIB) meant even fewer facilities would be able to "play the game." Those which did continue to operate (or which came into existence at the time) got their funding from the government - DOE, NSF, and others internationally - as the operational costs were now too high for any one facility to cover. And thus the User Facility was born.
User Facilities, like Oak Ridge National Lab, Argonne, Berkeley, Los Alamos or TRIUMF (or abroad, GANIL, GSI, CERN, and others), get their funding from government grants and thus are able to provide research capabilities to smaller groups who cannot otherwise afford it. Most places have a small permanent staff of administrative and technical support, and many have permanent staff scientists as well. But the idea of a User Facility is to be for "outside users," and so a system had to be put in place: hence what is commonly known as the "Program Advisory Committee," "Experimental Oversight Committee" or any number of other bureaucratic names. It's a committee, made up of both locals (from the facility) and renowned outside users (many from other facilities), to whom all proposals for experiments are submitted. The panel reads the proposals (which come from collaborations all over, including universities and smaller facilities, as well as internally), rates them on some kind of predetermined scale of scientific merit, feasibility and general coolness, and then sets the schedule for the next chunk of time (typically about a year). Sometimes, proposals are unsuccessful. Sometimes they are granted a trial run. Sometimes they're lucky, and are scheduled immediately.

But here's where the system begins to break down.

In theory, it works, and everyone is allocated time and experimental opportunity fairly, based on the same rubric. But in reality, because the facilities have grown so large and complex, it is difficult to know the experimental setups so intimately as was once the case. The limit of what is and is not feasible at any given facility is something that no one knows for sure, but that the insiders - the local staff and researchers at a facility - know much better than anyone writing a proposal from outside. And this becomes a hindrance.
If you don't know the ins and outs of a system, it is difficult to know what experiments might be simple and what others might be hopelessly complex. Conversely, if you are lucky enough (as I once was) to be one of the insiders, you are there all the time, you know the system and its quirks, you know what's possible and what isn't, and you're able to tweak any scientific justification with experimental capability such that your proposal is guaranteed to be more successful than one (perhaps even looking to study precisely the same reaction) submitted from outside. It is a fine system for those inside; you are still competing for time with other locals, meaning your science has to be as rigorous as possible. But for those outside, the system is faulty, because they must be scientifically rigorous as well as being absolutely certain of their (as yet nonexistent) experimental setup.
I have been fortunate enough to be on the inside, and to end up with my name on a myriad of papers because of it. But I have seen firsthand the effects of being on the outside as well. Consider the UK, where I now work: because of the dwindling funding of science over the last few decades, essentially no nuclear physics research facilities exist in the UK anymore. As such, one is pushed out of necessity to User Facilities beyond the borders, where one is always an outsider. Even when proposals are accepted - and they still are, for the system does work, if not perfectly - the damage is long term; the students and postdocs who actually do the data analysis suffer from lack of first-hand knowledge of the experiment and the facility (which in turn makes the analysis longer and more difficult), and often don't have the opportunity to participate in more than one experiment in an entire graduate career. Perhaps this is why such a large percentage of PhD graduates in nuclear physics in the UK go on to careers in industry instead of pursuing academia and research. It's difficult to find statistics, but anecdotally, about 50% fail to go on to posts in basic research. Conversely, nearly all of those we hosted as summer students at a national lab went on to join graduate programs.

So there are pros and cons, not all of which are easily quantified. With a new era of even larger User Facilities (like FRIB) dawning, these questions need to be tackled, not ignored. Will having ever more specialized, and ever fewer, facilities lead to an "us versus them" scenario in the end, with everyone fighting tooth and nail to be on the inside, that they might be able to do science? Or will the newer facilities be true to their spirit and actively provide equal opportunities to all researchers, whether locals or not?


  1. A model might be something like the Hubble telescope, where everyone is on the outside (I think). That would be the fair way to do it. If there is worthwhile insider knowledge, it really should be documented publicly. Indeed, there might be a user assistance team inside the facility that publishes adjustments, statistics, problems, etc. on a regular basis (They invented the internet, after all!). But doesn't do science.

    Yes, these issues crop up elsewhere. There was a huge problem getting protein crystallographers to publicly deposit their coordinates, being their crown jewels, as it were. The journals were complicit in that issue.

  2. There are typically liaisons who work with outside users, but they often are researchers as well. Most of the time, one finds that outside users who understand the dilemmas presented and are still keen to get experiments accepted will work with one or more of the local researchers on the proposal. You give up a little bit of your credit (by adding another name to the author list) in return for learning the subtleties of the facility and gaining better odds of success. Most people find that this is a trade that suits them, which is fine. There is still the problem of data analysis once the experiment is proposed and run; outsiders who take away their data and fight through it on their own encounter more troubles than someone familiar with the (entire) system. A common framework for all facilities would be one way of getting around at least part of this difficulty, but it would be an immense thing to have to actually implement.

  3. I find this post quite disheartening. The fusion community is going through this transition right now. While there have been several large User Facilities in existance, there have also been many $0.5-$1M-scale fusion projects at each major plasma-physics orientied universities. The word has come down from the DOE that this practice is now ending and "the fusion community will follow the successful model of the particle physics community" or else.

    To give an interesting perspective, it is viewed as a PR move rather than a fiscal one. Many labs working on small, disparate projects does not seem encouraging to the outside public. It is better to have one 'Right Answer' than multiple scientific approaches that question each other's validity, regardless of if it is the best.

  4. That is really fascinating, in (as you said) a disheartening way. I wonder if the real question is not about User Facilities, but rather "is Public Relations killing science?" While I've always been of the opinion that scientists who turn into popular science writers/TV personalities are no longer really scientists, it was on an individual level - if we, as the scientific community, bend our science to the public's opinion of what it should be, aren't we the ones ultimately responsible for science's demise? Have we all ceased to be scientists?

  5. With all due respect ... hold your horses right there. Fusion research has accomplished- what, exactly? From a public perspective, it has accomplished nada- nothing. While the dream lives on in further semi-absurd installations of ever-mounting complexity, the prospect of making something useful of fusion remains distant, if not a mirage.

    The public has the right to support science on a cost-benefit basis, in addition to the more basic academic support it gives to, say, English literature and philosophy departments. No power from fusion? You may be looking at a future as a theoretical rather than a practical field.

    As for popular science, it is essential that your funders know what is happening (productively, at least) in scientific fields. The connection between the public and science is carried on by those popularizers who have the social chops (and ambition) to span these worlds. Ivory tower folks may snoot down their noses at them, but their own funding depends upon them. The current crisis of climate understanding, and the prior crisis of evolutionary biology understanding stand as warning signs that if scientists do not bring the public along with well-pitched understandings of advanced scientific fields, scientists will be left howling into the wind.

  6. Burk, while I am as aware as anyone of how far away fusion energy is, I should remark that "fusion science" covers a larger range of research than merely tokamaks - plasma research in general usually falls under this title. However, you're right when it comes to science popularizers; we mustn't break that communication with the general public, as I've mentioned before.

    I wanted, however, to address your comment: The public has the right to support science on a cost-benefit basis, in addition to the more basic academic support it gives to, say, English literature and philosophy departments. Why should science fall outside of this "more basic academic support" category? Why shouldn't the public fund science for the sake of science? Must everything be a means to an end?

  7. Hi, Kelley-

    "Why shouldn't the public fund science for the sake of science?"

    Yes, very good question. It should and does, as you see with the LHC and Hubble. But then it doesn't, as in the case of Texas supercollider. The public is certainly receptive to scientific justifications alone. But if a line of inquiry is sold as a practical affair, as fusion has been, then it has to be judged on that basis as well. If the DOE thought that basic plasma research justified continuing all its small so-called "fusion" facilities, (or giving them to NSF), that would be one thing. But it doesn't, and I sympathize with that position, though I can't judge the details.

    If fusion researchers wished, they could drop the fusion rationale and defend their research on plasma grounds alone, as do other esoteric fields like condensate physics and high atomic number nuclear physics (I am straining my brain here). Or particle physics. Plasma physics in that case would probably be a relatively small field, not building monstrous tokamaks, etc.

    One of the most visible places this tension plays out is in space policy, where we balance the idiocy of "peopled" programs like the space station, which have zero scientific rationale and high PR value, with the real science of Cassini and Hubble, which have some PR value as long as there are popularizers to tell the story. The Mars rovers are something of a hybrid case. The withdrawal of Pluto as a planet was also an interesting case in the anals of popular vs real science.

    The same situation plagues the "war on cancer", which has only glimmers of positive results after decades of funding. There, I at least understand the details and appreciate that the basic science side of the equation has been very strong, laying the foundation for fundamental alterations in the practice of medicine ... sometime in the future. Thankfully, this future is quite a bit more discernable than that of fusion energy.

    So yes, everything must be a means to an end- the end that the project is premised on, whether basic or applied.

  8. Call manned spaceflight idiocy if you must, but I still get chills when I watch shuttle launches. When you ask little kids what they want to be when they grow up, they'll still say astronaut.


Think carefully before you post. I reserve the right to moderate any comments posted to my blog.