Previous posts by fellow OT blogger Jenny Rohn (here and here) and me (here) about fellowship funding have sparked a lot of debate and interest. With the report that Science is Vital has just produced for David Willetts described in another of her posts, plus the round table that Jenny and I will both be participating in later in the month with Willetts, Paul Nurse and others at the Royal Society, it seems timely to revisit this issue. More specifically, I’d like to revisit the topic building on the post that Tom Hartley wrote as a follow up to our original ones, about what might represent ‘fairness’.
Let me start out by saying that I have every sympathy with those early career researchers who feel immensely frustrated, not to mention depressed, by the current system. There are a huge number who end up being disappointed, many of whom no doubt could have thrived if only they had been given the right opportunity, and who would have given much back to the taxpayer who will have contributed to their education and training up to that point. But I still don’t like the use of the word ‘waste’ when these people are discussed, because education is (in my view) a public good, and these people – as long as they don’t end up so embittered they give up any attempt to use the skills they have learned along the way by opting out completely – are likely to continue to contribute massively to the economy in many different ways. Becoming a school teacher, a festival organiser or a freelance writer (just to name some random roles they might assume) may not equate to their dream of continuing as a bench scientist let alone a PI, but I don’t believe it amounts to a waste of their education and training. Furthermore, if the funding system made it possible to create more permanent postdoc positions I would be absolutely delighted. These roles are invaluable for the individual and for their PI (and do exist, as far as I can see, in some STFC funded programmes). But such positions cannot solve the fundamental problem because there simply wouldn’t be money to create sufficient posts to fulfil the aspirations of all.
Tom Hartley felt that the funders – from whose point of view my earlier post was in essence written – are not really “doing the best that can be done with the money available”, or at least that he didn’t feel at all confident that this was so. He is intending to do some modelling to check this out, but I am not aware he has yet put the results out on the web. In offline comments to me he indicated that it might be much ‘fairer’ for instance, if the money was spread more widely and more thinly, so that luck played a less important role in people’s careers. So let me tackle that point of view, by raising the question of fairness for whom? This I think gets to the heart of why modelling may be so hard to make meaningful.
Suppose a fellowship competition changes its processes, so that instead of awarding 20 x 4 year fellowships it offers 40 x 2 year fellowships, will this make things fairer and more popular? No doubt it will feel so to the 20 who get a fellowship now who wouldn’t have under the earlier rules. But actually it means not one of these people will be able to take on a PhD student, or apply for most grants. The time is too short. I am aware that people currently with 3 or even 4 year fellowships worry about this shortness. Everyone wants fellowships to run for 5 years so that they have time to demonstrate that they can supervise students and win grants. So I am not convinced Tom’s suggestion on this front would actually make the postdoc population at large feel more content. Of course none of the 40 putative fellows would know whether they would have been in the 20 who would otherwise have got 4 year fellowships and so ought to feel aggrieved, or in the additional 20 who have now got something they would otherwise not have had and ought to rejoice the system had changed.
So it seems to me there are 3 measures of fairness of outcome which are likely to be incompatible. Fairness to the pool in general – which is no doubt what Jenny and Tom focus on; fairness to those who succeed in one fellowship competition and want to demonstrate how much they can go on to build on this success and generate more fantastic results; and fairness to the taxpayers who want to know their money is being wisely spent on top-notch candidates without massive associated overheads due to having an impossibly large number of applicants (with the concomitant nightmare of heavy administration and multiple selection panels), if all eligibility criteria were removed. The latter scenario, I will stress again as I did in my first post, would not lead to the removal of luck, it would just manifest itself in different ways.
An alternative suggestion to spread the money more evenly would be to say, if you were nearly successful in one round you were given bonus points for the next, whereas if you were successful you were totally excluded – this is an approach that seems to be implicit in Tom’s post as a way of levelling the playing field. What this means is the truly excellent – those who do have the skills and not just luck – will be hampered to make way for others who may be less good. The bone of contention is that these others may in fact not be less good, merely less fortunate or less well-positioned. Funders need to go on the evidence put before them, as scientists surely should. And if the evidence is made up of a mixture of luck and skill any selection panel can only go so far in disentangling the two. This is in essence the same challenge that universities have in widening access: go to the ‘right’ school, have the ‘right’ parents or happen to be lucky enough to have a brilliant teacher and life is much easier than if these things aren’t present in your life. My university works really hard to try to distinguish applicants who have been well coached but are merely solid from those who haven’t had the advantages of such preparation but are inherently smarter. It is not such a very different problem and is equally hard to resolve. Contextual data, as with university admissions, may or may not be readily available or usefully interpreted by selection panels trying to work out just how ‘lucky’ someone may have been.
Science is Vital have made the point to David Willetts in their submission that losing so many postdocs after many years’ experience gained because they haven’t made PI is like throwing out all teachers who don’t get to be Heads by 40. The analogy is not perfect, though, because there is far more turnover in head teachers than in Lecturers and far fewer people chasing the positions. So the supply and demand pattern is totally different. Indeed in science teaching there is a serious shortage in many subjects, such as my own of physics, so the parallel is not at all close. As long as we have a huge number of aspiring PI’s who are willing to take the risk of there being no opening of a permanent position for them in the end, the problem of the system ‘rejecting’ large numbers is bound to continue. The simple supply and demand solution would be to have fewer entering at the bottom as lowly postdocs. If there were fewer of these, there would also be more money to create permanent senior postdoc positions. Perhaps that is the logical solution to this problem, rather than homing in on the specifics of a particular fellowship scheme’s eligibility criteria. However, it isn’t clear to me that that is any fairer, since the rejection process will just occur much earlier in one’s career, where the evidence is bound to be even less concrete: how can you judge someone’s innate brilliance infallibly at the end of a PhD? That is a time when there will have been extremely limited scope to demonstrate independence of thought, leadership, originality or any of the other qualities that ultimately matter for a successful career.
So, reforming the postdoc career structure is clearly something that needs to be thought carefully about. Undoubtedly PI’s should do a better job of informing those starting out of the risks inherent in taking on a postdoc position and ensure that the skills learnt include many generic/transferable ones which the postdoc can usefully (and enjoyably) take if they leave the laboratory. But, with finite money and far more researchers dreaming of running their own teams than positions available, it is inevitable there will be disappointments. Shaking up eligibility criteria for fellowships will never solve the fundamental problem. Nor can creating a few permanent positions, desirable though that may be; it will only create a different kind of lucky elite.
After I drafted this, I came across a post about the pros of multiple short-term contracts, written by Tom Webb, a postdoc who has had 4 successive 1 year positions. It provides an interesting contrast to the usual unhappiness with that aspect of the system, but it would take my own post too far in a different direction for me to be able to address that question here.