I had been writing this post, but had put it aside (honest). Yesterday Jenny posted about the same topic, so I’ll take a break from putting envelopes around trees, and finish what I was writing.
If you have been hanging around in the scientific blogosphere for some time, you get to read about several people who have been through the scientific mill, doing PhDs, and perhaps post-docs, but at some point finding they have to leave science. For some it is because they have fallen out of love with academic research, for others it is because of difficulties finding their next job. This is the problem that Jenny was worrying about.
As a PI I also see the other side: the pressure to acquire funding for research. In practice, this means getting bodies into the lab or in front of the computer. These will either be PhD students or post-docs, because they’re cheap. Students in particular. A large number of junior posts are thus created, but senior posts are few and far between. My impression is that this has got worse over the last 20 or 30 years. How did it get this way, and what can we do about it? I can speculate on the first, and wave my arms on the rest. And, this being a blog, I will simplify hopelessly. So be warned.
The problem, I think, is that the higher education system has evolved. Back in the old days (according to the ancient myths), the purpose of a PhD was to train the student to be a researcher: an apprenticeship in academia. Most students would expect to find a job in universities, or perhaps another government research institution or in industrial research.
But at some point, research became more professional1. This was linked to government funding: putting money into science because it was Good For The Country. A lot of the money was soft, so the most convenient way of using it was to hire students and post-docs. The effect was, of course, to increase the number of PhDs produced and Post-docs employed, without necessarily increasing the number of jobs available. Most academic research is carried out by universities2, who’s permanent funding usually comes through education ministries rather than the science ministries who give out research money. Between this and the short-term decisions about science funding, we have a recipe for inconsistency. If money is given to research councils on a short-term basis, it will be used to train PhD students, who then can’t find the permanent positions, as they are largely paid from the education budget (yes, this is horribly simplified).
Of course, the same thing happens at all levels: a lot of undergrads don’t have a career in the subject they studied for their degree3. But it is accepted that an undergraduate degree provides a more general education, so that one can go into a wide range of jobs: research, teaching in a school, managing a nature reserve, working in the city, whatever. But, because of its history as an academic apprenticeship, the PhD is seen more specifically as a step on the path of a research career, and mainly in academia to boot.
About here is where Leslie steps in. I must admit I know very little about Leslie as a person, and can’t be bothered to flex my google-foo. But he formalised a method that can be traced (via Fisher) to Euler and perhaps elsewhere4. The idea came from happy acturists, who were looking at changes in the human population. What they had done was to create tables of rates of survival and reproduction for different groups of people (coal miners, bishops etc.). With these, one can work out the change in population size, by solving one of Euler’s equations. With this framework we can estimate things like the stable population size given rates of attrition, how the population will change.
Where Leslie and his matrices enter into all this is that we can think of progression through the academic system in the same terms. Leslie matrices use discrete classes of individuals, so we could use the following categories: undergrad, masters, PhD student, post-doc, lecturer, professor, fossil. For each class there is a probability that any individual progresses to the following class. If they don’t they “die”: they become an accountant or perhaps work for a journal5. If we know the rates of entry and fall-out, we can calculate the numbers in each class. Now, if we could restrict entry and let everyone do what they want, we would end up with a nice, distribution if happy people: the drop-outs would leave because they realised that research wasn’t for them6.
But, unlike the systems Leslie (and Fisher) looked at, the academic system is regulated by external economics. There is a ready demand of fools students wanting to do a PhD, and the number that can be accepted into slavery bondage our exulted ranks is determined by the amount of money the government has to throw at the research councils. Once they get their PhDs,what do they do? The system is geared towards them doing a post-doc, but that depends on the research councils now having the money to hiring them. And then many post-docs want a job as a lecturer/senior scientist. The money for that depends on long-term strategies, and to a decent extent on the education (not research) needs of society. So the drop-out is determined by what we can afford, not what people want to do. Hence the stress and sadness.
I think it’s clear that this crates a problem, but what to do about it? I think there are two general strategies that can be used. The first is the one Jenny was thinking of: reducing the intake of PhD students. Jenny’s suggestion of limiting the total number of students is one way of doing it, but it would take a lot of organisation and even more paperwork (I’ve supervised 2 students in Finland, so who will follow that now I’m in Germany?). Quite frankly, it’s too much hassle. And how does one, as a PI, decide when to pick your student? Get them early, and then ‘retire’ (just when you’ve worked out how to train students). Or leave it until later? And the good teachers are stopped from training once they’ve hit their quota.
A better approach might be to restrict the number of PhDs that can be registered for training in any year. This is easier to organise (much less bureaucratic memory needed), and also means that there can be competition so that the best teachers can train more students (OK, this is somewhat idealistic).
The problem with this is how one get the extra science done. Jenny wants to have more permanent junior positions. This is possible, but I think there are two problems. The obvious one is the cost effectiveness. This would need data to decide upon: I’ll leave it for there. The second problem is that it changes the structure of science careers: we would have to create the whole career structure for this new strand of scientists. I guess the Civil Service’s system of Scientific Officer/Senior Scientific Officer etc. could be adapted. The far-reaching effect of creating a new strand of scientist could be interesting: a new class system would be created.
The alternative strategy is to allow more junior researchers to leave research. We have to realize that a large proportion of students leave the university system, for example going into industry (think BIG EVIL PHARMA). But many will leave research. If we accept this, then we have to re-position the PhD as a more general education. This means training students in more than just research skills (a couple of people on Jenny’s thread mentioned that this was being done). But it also means cultural shifts, with academic research not being perceived as the acme of professional development. We, as senior academics, have to be happy to be training students (and post-docs) for a life not just beyond academia, but beyond research.
The questions I’m throwing out are not things I can answer alone, obviously. They get to some of the deeper parts of the scientific culture, and how it has evolved. The more I’m in science, the more I’m aware of the tension between the old culture, where a scientific field was small and (in many ways) amateur, and the newer professional science we have now. Science has to adapt, and the culture is one that in many ways seems conservative (how many of you had to write your theses as a monograph, rather than as a collection of papers?). I’m sure we’ll muddle through, but only because of the difficulty in changing the culture by diktat.
In the mean time, we still have a duty of care to our students and post-docs to ensure that they are not just left on the sidewalk pavement, but are helped to find a suitable job and career. Ha! Cheap words and good intentions – I hope I can live up to them.
1 I think this is largely a post-war phenomenon. The whole professionalisation of science looks fascinating, and seems ripe subject for a semi-academic book. Perhaps someone has written one.
2 He writes, having been employed by 3 research institutes and only one university: and for a time was in a non-teaching institute in the university.
3 Cue jokes about English major flipping burgers.
4 I haven’t followed up the history of this.
5 Although I know of a couple of journal editors who were resurrected back into academia.
6 This would include my brother, who did his PhD on reclaiming coal from coal tips – something my mum, the daughter of a Yorkshire coal miner, found hilarious – and now works developing UK passports.
A peerless post, Bob. But when you wrote ‘Leslie’ I thought you meant one of these…
I’ve been brought here under false pretences.
Great post, Bob – I didn’t know about Leslie but it’s great that someone else has thought about this in terms of population biology – someone who actually can do maths, unlike me!
I want to stress that in my dream, Phase I, I said that PIs would train far fewer replacements. But I wasn’t stipulating the mechanism – your idea of centrally restricting them (as happens in undergrad places) is an excellent way to tighten up the ingress. I don’t think it’s feasible either for PIs to choose when to stop reproducing – if there is an equilibrium to nurture, then this decision must come from exterior beancounters who are keeping track.
As I just cross-posted on my own comment thread, I’m disappointed that you seem to favor a solution that offers permanent scientific positions only to the PI – business as usual with a bit more career advice for trainees. Restricting the number of students who can start is a good idea, but based on my dealings with students (and youth in general), too many will still think they can be lab heads, and will desire this outcome, no matter how hard you try to convince them it’s a slim chance. I also think it truly is possible to employ professional junior researchers without blowing the budget: you would need a smaller lab to get the job done if half of its members, say, were highly experienced. So just saying “it’s too expensive” without at least considering the economics and various models seems a bit of a cop-out.
Regarding the ideas about restricting the numbers of PhD students: there’s actually a trend in the opposite direction at many US universities, at least in the biomedical sciences. Graduate schools rearrange their PhD programs (e.g. “integrating” the application and intake process, as well as teaching, across several departments), increase spending and effort on marketing, form alliances with other local universities, etc. – all in the name of increasing the numbers of incoming grad students. Population dynamics (the baby boom of the Baby Boomers), economics, interest in studying abroad, and probably some other factors that I haven’t thought of, increase the numbers of students who apply to graduate programs. Right now it can be tough for a young person to find a job with just a science bachelor’s degree in hand, and if they’re not suited for an MD or DDS program (at the end of which there are indeed excellent job prospects), a PhD program might seem like a good option.
[/anecdote warning] I was just at a local meeting here, and heard a talk by a former colleague who moved his lab to Ireland. From the obligatory lab photo slide at the end of his talk, it seems that he has many more lab personnel and students in his new location than he did here. Not sure whether that’s simply a function of his new position, or whether there are more available research positions in Irish universities. The type of research he does appears to be essentially the same, in terms of labor-intensiveness per publication unit.
Thanks for your comments!
Jenny – I didn’t actually make any comment about permanence in the “super-lab tech” class of people: the career structure was too much detail, and too speculative. I agree that more permanent jobs is better, but I’m not convinced that having more senior staff on PI salaries is going to fly. More permanent technicians will work, because their salaries are cheaper (I’d prefer everyone to have higher salaries, but alas economic realities are real). Beyond that, I think we’d have to see some costings for the different models, and the scientific and educational outputs.
Kristi – Yes, that’s what I see in Europe as well. It’s all wrong, unless we put the production of PhD students into a bigger educational pipeline.
OK Bob – sorry not to be clear. In my dream, the permanent staff are necessarily on a salary structure lower than PIs. An experienced postdoc in the UK currently ceilings out at about 40K, so I wouldn’t expect much more than 40-50K. What you lose in salary, you gain in permanence – and not having to spend all your time in an office!
The posts by you and Jenny have voiced what myself and my three fellow 6th-year (yikes!) PhD students in my lab have been thinking about for the past few years. Once we had all passed our prelims and it seemed feasible we would eventually get our degrees, we started to look a lot harder at the prospects of becoming a PI.
The encouragement we received as undergrads and 1-2nd year grad students of becoming PIs at big universities with productive, grad-student labs seemed not so much as the god-given right of a PhD as it had been “sold” to us. Like you and Jenny recently discussed, the numbers just didn’t add up.
However, one career option that many of my peers having been looking into that I don’t remember being mentioned it still becoming a PI, but not at a “tier 1” university. Instead, having a lab and carrying out research at a primarily undergraduate institution. There’s a lot of these schools in the U.S., where grad students and postdocs are few and far between. You don’t always need to do a postdoc to get one of these PI jobs (but its becoming more common). Yes, you’ll still need to secure funding and get papers out, but the expectations are a little different since you lab will mostly be undergrads doing senior theses and maybe a tech or postdoc sprinkled in here and there. The salary is quite less than a PI at the big research universities, but maybe this fills some of that gap between postdoc and typical PI that Jenny talked about?
Jenny – I guess that would work, although it could be seen as a dead end position, with little chance of advancement.
Elizabeth – I can’t really comment on the US system, although I guess that a lot of post-docs do take this route. As I understand it, these positions are primarily for teaching, so again the number of available jobs is determined by educational needs, not scientific.
Perhaps e should view the PhD as a teaching degree. 🙂
Nice post Bob. When I speak at new postdoc orientation I try and explain to the n00bs that they need to accept their fate is uncertain, and that they need to train broadly and confidently for a varied future.
Most got it after a bit.
And then our postdoc Office stopped inviting me to orientation. I was a bad influence and was scaring the little darlings. That’s part of the problem, the utter naivety of the audience. next blog post on these lines coming soon >:(
Thanks. I enjoyed your blog post too (ooh, what a little love-fest!).
The discussion here and at Jenny’s post is fabulous.
I think that a really key part of the problem is to reduce the number of PhDs graduating from any given university. But I wonder if funding bodies could also create some longer term grants or operating grants. In the arts, groups start off on project to project funding but after a while they get switched to operating grants. If lab groups could get “operational” funds then, permanent research positions, as was suggested could be funded.
@Elizabeth Moritz Actually, these institutions are experiencing a huge increase in the number of applicants. A friend of mine got a job interview at one of these small teaching/research schools in the US and it had 500 applicants. Of those ~50 could easily have done the job i.e. postdocs with a number of high level publications. I think it’s a myth to think that those schools will fill the gap. I’m sorry I don’t have numbers on those figures.
Elizabeth MoritzDilara AllyThese kind of “mainly teaching” US schools actually pay noticeably better than the big research Univs in the UK…!
Almost all UK research Univs actually represent a compromise between what a US-based reader would recognise as “research” and “teaching” schools; Faculty teaching loads would almost never be as low as in the big US research schools, and the teaching load “titrates upward” as one’s grant income lessens. So different “lab PI” Faculty may have teaching loads that differ by four to five-fold, and nowadays there are also teaching only posts there, as discussed a bit on Jenny’s blog.
Agree with Dilara that a bulge in well-qualified people looking for jobs in such US “non tier 1” institutions is likely to raise the bar for employment there. So while they may take up some of the slack, I can’t see it solving the “decade of postdoc-ing slog then what?” problem.
Thanks for your comments, Dilara and Austin. It doesn’t surprise me that teaching institutions are getting tonnes of applicants: it’s the obvious place to spill over.
We keep on coming back to the point that a PhD is about research, but a lot of the jobs involve teaching.
@Bob
Lovely post. I like the “different scenarios” idea you paint. I guess my main question is (putting the idea that the people who get a PhD and sometimes do a post doc wants to leave science) “will this structure be cost effective?”. After all, many of the universities in Europe at least are publicly funded to some degrees.
I agree with what’s been said on Jenny’s side about the criteria of having a post doc/PhD in order to get a job, it’s not that you really need it for the position offered but it makes it easier to choose between the candidates.
Maybe what I would’ve like the most if I had to redo my PhD would be to add on some “industrial skills” (i.e. budget making or other mesurable things that is not lab bench or writing papers based). Then again, I don’t think I will ever redo my PhD 😉