What has changed in science and what must change – I. Rethink the scientific career.

I have been pointed today by my friend and brilliant scientist Giorgio Gilestro to an interesting blog post of his on how science should change. My reaction was “This is really really interesting and raises a lot of fundamental issues; yet I disagree almost entirely.” However 1)it’s really good food for thought and 2)you can’t understand the post if you don’t read it, so, well, please read it now before continuing (I will quote relevant parts to which I answer but, still, makes more sense to read it).

Ready? Ok, here is my first part of responses. More will follow: the points Giorgo raises are many and, well, synthesis is not my strength.

We want group managers and we want to acknowledge them.

Giorgio’s post reflects a common but mistaken frustration of the young researcher: the apparently “cerimonial” group leader role and the fury at him getting an undeserved merit. He writes:

[PhD students or postdocs] are the ones who do the work, without doubts. In many cases, they are the ones who do the entire work not only without their boss, but even despite the boss. […] The principal investigator [..] will spend all of their time writing grants to be funded, speaking at conferences about work they didn’t do, writing or merely signing papers. […]

Let us not bother with the scientific contribution of a group leader (which often, even if subtle, is not negligible). Just the fact that there is a guy that writes grants to get funding to give me resources to make my job should be fully recognized. Writing grant applications is one of the worst chores of the scientist routine. This alone makes the role of group leader fundamental: there is someone who takes care of funding (and all the PR along with it) while I can do work. And this someone, by funding a group, can fund stuff that a single person or small group of students would never realistically have (does every biophysicist have to apply to fund purchase of his own personal 900 MHz NMR? Think about it while I laugh.) I am already overwhelmed by the fact of having to write (lots of) applications to get a wage. We can’t expect a scientist to micromanage everything: this is simply not humanly possible.

Also, going to conferences presenting your work is not something to despise: they’re making advertising of your stuff, after all. You can do it yourself, for sure, but chances are that he has already the right personal connections, that he knows better than you how to do the PR spin etc.etc. -and in any case, advertising rarely hurts. They can also put it in better context with the rest of the group work. The same for review writing and paper writing.

Everybody knows that it’s the Ph.D. or postdoc as first 1-2 authors who did most of the job, so the issue of having the group leaders signing your papers seems moot to me. The recent timid trend of introducing statements of the effective contributions in scientific papers should only help in making clear what is already obvious.

I could write about the fact that groups and individual benefit from a “directors’ cut” -good group leaders give a vision and a framework to an individual’s work which helps put it in context and foster collaboration with your closest collegues- but let’s not bother. The point is that, practically, science needs manager roles. And that’s what baffles me when Giorgio writes:

The system as it is does not reward good scientists, it rewards good managers. You can exploit creativity of the people working for you and be succesful enough to keep receiving money and be recognized as a leader but you are feeding a rotten process.

The process is not rotten (from this point of view): it rewards good managers with management roles because, er, they have to be in management roles. It is complete nonsense to have a system needing managers that puts brilliant non-managers there. And here we arrive close to the crux of the problem.

Group leader means good scientist, but good scientist does not mean group leader.

A group leader has to have been a working scientist, and possibly a very good one -after all, you can’t manage something you don’t grasp brilliantly, especially in science. But -and that is the crucial issue of all this first part- not all scientists are meant to be group leaders. And that is one of the fundamentally broken parts of the system: the scientific career is meant as a forced progression from dependent bench work, to independent bench work, to management role. And this is what is wrong.

Giorgio correctly points that:

The number of people entering science grows every year, especially in the life sciences. The number of academic position and the funding extent is far from being sufficient to cover current needs. In fact, about 1-2 in 10 postdoc will manage to find a job as professor

The ratio is absolutely correct per se: it’s reasonable to think at only 10-20% of postdocs is suited to a management role. The problem is the waste of the other 80%. Let’s leave aside the bottom 30% in a hypothetical quality scale, which is probably good for some other job. This makes 50% of decent-to-pretty good postdocs, being thrown away and substituted by new ones. Personal suffering aside: does it make sense?

Enter the Eternal Postdoc

Giorgio makes a proposal which I find the creepiest of all his post:

In gg-land, the imaginary nation I am supreme emperor of […] there are no postdocs as we know them. Labs have students who learn what it means to do science. After those 3-5 years either you are ready to take the leap and do your stuff by yourself or you’ll never be ready anyway.

And why being ready at all? This kind of proposal implicitly assumes what is one of the most puzzling aspects of science nowadays. As said above, scientists are thought to either become independent investigators (that is, group leaders), or to go away.

This has an interesting first drawback: As of today, science is nearly always made by people who have less than 10 years experience in their skill sets.

This would be bad enough for a single kind of skill -mastering a single musical instrument takes more time already- but science requires an awful amount of very complex practical skills. There are single instruments, like NMR spectrometers or AFM microscopes, you practically never completely master in a normal career: you perhaps learn to use reasonably well the fundamental subsets of their potential, but mastering a technique from top to bottom is often out of the question. I leave to the reader to imagine the utter waste that this means.

I find it totally absurd for an enterprise -and especially one as science- to throw away its workers just after 3-5 years of them learning their job. No sane technological company would sack their engineers after only 3,5 or even 10 years to take up only new, pristine, inexperienced ones: yet that’s exactly what is proposed.

Why doesn’t this sound as crazy as it should be to the scientists’ ears? Because of the assumption above: that a scientific career automatically means getting more and more into independent, management roles.

I challenge this. I think science to be efficient has a great need of people which never take the independent leap, for the simple reasons that, as much as brilliant thinkers, efficient managers and visionary leaders, science needs people who simply know how to do their job. Bertrand Russell said:

“No transcendent ability is required in order to make useful discoveries in science; the edifice of science needs its masons, bricklayers and common labourers as well as its foremen, master-builders, and architects”[1]

That is why in my imaginary nation, the true backbone of the groups would be eternal postdocs: people who maybe are not as creative and deep as their leaders, but that are good at doing benchwork, designing experiments, know their systems from the inside-out, can take the time of learning new techniques in full and to maintain tools developed in house. People who do not live with the anxiety of getting a position or die. People who would not be simple technicians, people that have their own projects, fitted into the group framework, just like a Ph.D. or a young postdoc. People who maybe are damn good scientists, but that simply don’t want or are not able to manage grants, devise new research projects and be their own advertising agency[2]. They just want to do their research.

Apart from the boost in human efficiency for a research group, such a move would finally open the possibility of a true lifelong career doing just science: something which now is simply impossible. People who want to be managing research could apply for the managing positions, while people who just enjoy research without the hassles could do that.

No free lunch

And here it is where more research money would be essentially well spent: you have to fund the young scientists who still don’t know what to do and the “eternal postdocs” -that is, the permanent bench working personnel of research. First of all, let’s throw away the canard that research money is destined to be scarce. Research money is scarce, but it is so mostly because it is politically a badly beaten and forgotten Cinderella. Try to find research money in this graph (hint: it’s yellow) and let’s move on.

So resources should be larger indeed, and a good part of them should be spent in guaranteeing groups core of permanent or semi-permanent personnel (of course subject to review, like in any organization) that keeps the day-to-day work going smoothly and skillfully. Young brilliant students can and will continue to pursue the exciting research projects that they devise along with their group leaders; new group leaders will arise as usual. But it’s no more “either you climb the ladder or you die”: there would be now the option of becoming, if you’re skilled in your work and useful for a lab, a respectable science worker, of being, to say it differently, a scientist forever.

Of course such a world puts on the table the problem of how to recognize and evaluate the job of such people. They would publish papers -a majority of first names could be easily theirs- but I feel that this has to be more relevant for the leader than for the researcher. My proposal is simple: it’s the group leader who decides. The eternal postdocs must not be bothered with grant applications and funding chores: that’s the job of the group leader or the ambitious young wanna-bes. It’s the managers who find the money and therefore it’s up to them to evaluate them. It’s simply in the managers’ best interest to find a good group of permanents to work on, to sack the bad ones and recruit the good ones, because otherwise managers’ results -and thus funding- will drain away. It would be like in any other enterprise, no more no less: good managers thrive because they hire good people and know how to make them work.

Eternal postdocs would not be hired on the basis of being able to “sell themselves”, of being their own micromanagers and being measured by some awkward sciencemetric (more on this next time): but simply on the basis of being known to be skilled and trustworthy at the scientists’ job by their managers.

Who spends more, spends less

On a sidenote, I find extremly weird Giorgio’s proposition on “low funding is better”

Another fun thing about doing science with less resource is that you really have to think more than twice about what you need and spend your money more wisely. Think of the difference between buying your own bike or building one from scratch. You may start pedaling first if you buy one, but only in the second case you will have a chance to build a bike that run faster and better. On the long run, you may well win the race, I believe (of course you should never reinvent the wheel; it’s OK to buy those).

This sounds really romantic but it’s nonsense. There is a reason people buy bikes and not pieces of metal and rubber: efficient division of work. Michael Schumacher doesn’t build Ferrari cars. It’s simply hideously inefficient and slow to rebuild your bike each time you have to ride a race. If you are such a good bike engineer that can build a faster and better bike than the ones on the market, chances are that you have spent a lot of time and resources in becoming a bike engineer: and this doesn’t only mean that your job is to be a bike engineer rather than a biker, but also means you spent, indirectly, much more money as time. If you really are a biker and you want a faster bike, pay the people who know how to do it.

And here again, the division of work and maintainance of in-house skills comes back. Relying on new students that have to learn techniques again and again is like having to learn to rebuild a bike from scratch for every race, instead of paying someone who already knows very well how to build one.

An ethical perspective

Finally, I must say that my proposal also has an inner ethical foundation. One of the problem of science is that it is a psychologically exhausting job. You not only face the perspective of being in competition with lots of smart people and therefore facing the choice of either feeling overwhelmed and diminished by comparing yourself with competition or, conversely, becoming a self-righteous smug smartass (or both). Once you finish your Ph.D. you basically are in the situation that you will never know where you will be next year, and in the situation where you will have to literally humiliate yourself and your research in “selling” it to get funds, wasting time knowing that four out of five of your applications, if you’re lucky, will be rejected.

There is people who is happy with that. But this means that today, often, it’s not the more brilliant people that become leaders, but the most resilient: the ones that will carry on stubbornly without breaking down. A lot of brilliant people simply throw the towel before. This is a huge shame and a huge problem for the research community. Since I’ve been in Cambridge I’ve suffered a huge nervous breakdown that has put me on antidepressants for the first time in my life, and I’ve seen practically all my collegues, regardless of their background, suffering similar things. I’ve seen people literally on the brink of suicide. And sometimes people do it..

Creating a backbone of skilled scientists which are employed semi-permanently and permanently would not only avoid a practical waste of human resources, but would also give another long-term career opportunity for the less ambitious but nonetheless skilled and enthusiastic scientists. It would also make a lot of people feel at ease, being in the right place doing the job they love, without having to fight with overcompetitive hawks.

Ok, this is the end of rant 1. There are a lot of other points I want to touch, and a lot of ramifications of this, but for now, that’s it. I hope this is the beginning of a conversation.

Footonotes

[1]B.Russell, The Place of Science in a Liberal Education

[2] Example: As a biophysicist-turned-bioinformatician the first thing that jumps to my mind is building, documenting and maintaining scientific software. This is an highly skilled, creative, complex and long-term job which nowadays is mostly made in their spare time by researchers doing research. This has a lot of practical consequences: scientific software packages are often poorly maintained, hastily coded, poorly usable and often forgotten after a short period. The point is obvious: either you build a strong community around a tool, or once the Ph.D. who developed it published it, he/she simply can’t get around it anymore: there’s new papers to churn and new stuff to do. (I know that: I am guilty of this). Why isn’t there people paid, long term, just to do that? (Actually sometimes there is, for example to maintain crucial community databases, but it’s the exception and not the rule)