Rambles around computer science
Diverting trains of thought, wasting precious time
Mon, 02 Dec 2019
Postdoc myths
[I'm on strike at the moment, largely in solidarity with my more precariously employed colleagues, whether hourly-paid or fixed-term or never-endingly “at risk of redundancy”. So it seemed a good time finally to finish and publish this post. I wrote most of it during the final couple of years of my seven as a postdoc, which ended in 2018.]
Lots of things are said, written and believed about postdoctoral researchers that are simply not true. This matters because real policies, initiatives, attitudes and actions are shaped by what people believe—true or otherwise. In this post, I'll tackle a depressingly long list of such myths. (I'm trying to keep this post snappy, but the flip side is that I have left out examples in many cases. For some, I also have hard data. So let me know if you'd like more specifics on anything.)
Myth: postdocs formally exist. In almost all universities I know, formally there is no such thing as a postdoc. In research councils' view of the world, it's the same: there are no postdocs, only “Research Assistants” and “academic staff (lecturer or equivalent)”. This matters because when practice on the ground no longer matches the ontologies on paper, systems become prone to poor outcomes and to abuse.
Myth: postdocs are homogeneous. Generalisations and stereotypes abound, both in writing about postdocs and in commonly held beliefs. This is unfortunate because postdocs are a highly heterogeneous bunch. Lumping them all together encourages wrong stereotypes. When these stereotypes hold sway with funders, institutions and departments, misguided policies result.
Myth: postdocs are all aspiring academics (lecturers). Clearly, some are. But there are many skill sets required in a healthy research environment. If you agree with “required”, then it follows that there should be a career path for all of them. Although there should be, currently there isn't. Once upon a time, the University of Cambridge did embrace this idea and had a variety of offices which had real status within the university, with titles including Senior Technical Officer and Computer Officer, as well as the research-oriented Senior Assistant in Research and Assistant Director of Research. These practices are mostly forgotten, and these posts replaced with lower-status unestablished positions: on the academic side, “Research Associate” is increasingly a catch-all, while on the other, technical officers are far fewer and computer officers are no longer the “peers of academics” that they once were.
Myth: postdoctoral work is “study” or “training”. It isn't; it's actually doing the actual research. I had to grit my teeth when applying for some funding that mentioned “postdoctoral students” in its particulars. Meanwhile, at a publication venue I plan to submit to, conflict-of-interest rules mentioning “advisor/advisee” seem to think there is a “postdoc” version of that. There isn't. At any career stage, we have people we turn to for advice, and people we work with. But by definition, someone with a PhD is a qualified researcher, not a student.
Myth: postdocs are “intermediate” between graduate students and more senior positions like research fellows and academics. The phrase “postdocs and PhD students” abounds. But in a university with many postdocs, the population of Research Associates is on average older and has more research experience than the holders of many flavours of early-career research fellowship. That's not surprising when the latter positions come with time limits (e.g. years since PhD) whereas the former don't. People can be postdoccing well into their thirties, forties and sometimes beyond. The “overgrown graduate students” caricature is wrong, disrespectful and leads to wrong-headed policies. (For a game of bingo, try this New York Times article from a few years ago.) According to University of Cambridge data current on 30th November 2017, of university-payrolled Research Associates and similar, over 40% had more than three years' service in the role, and around 10% of the total had over ten years of service. These numbers are underestimates of post-PhD research experience because they exclude postdoctoral experience at other institutions, and because the “and similar” positions include some of the aforementioned research fellowships which lower the average.
Myth: postdocs are on a journey to “research independence” (but are not there yet). This line is popular with funders, who don't seem to realise that their cause and effect are backwards. “Independence” is in practice a statement of one's funding status, not one's stage of personal development. As the mix of funding, in the UK and elsewhere, has moved further and further in favour of project-based grants, and away from institutional funding, hey presto! We have fewer “independent” researchers—on paper, but not in reality. In the UK, suppressing “independent” status is also a useful tactic for gaming the REF, as long as postdocs always co-author with their PIs. (If they don't, their publications are curiously lost into the REF-ether.) Again, the “paper ontologies” are a poor caricature of reality.
Myth: the opposite of “postdoc position” is “permanent position”. This comes up time and time again, but is completely false, at least in the UK. In all institutions I know of, academics (i.e. “lecturers or equivalent”, to borrow an RCUK phrase) may be appointed on limited tenure. They remain first-class citizens for the purposes I've been describing. Yet the justification for depriving postdocs of any given right or privilege is usually “they're not permanent” (a line often pulled out on-the-hoof, rather than reflecting any real rule). In fact, many postdocs are permanent, legally speaking, thanks to the 1999 EU Directive on Fixed-Term Work. Even those who aren't have a legal right not to be treated less favourably. Sneaky tricks skirting or infringing the edges of these laws are among the many ruses used by universities to keep their research staff dangling.
Myth: postdocs are itinerant, unlikely to be at the University in a few years' time, and/or are otherwise “not committed” to the university. To the extent that this is true, it is circular: funders' policies, and institutions' interpretations of them, base themselves on the assumption that postdocs will move on, and proceed to help make that assumption true. In Cambridge earlier this year, a certain fly-sheet had the temerity to claim that Research Associates did not deserve representation because they had not shown “commitment” to the institution (and that the university was not at all complicit in the underlying funding changes that had precipitated the growth in postdoc numbers; no, certainly not). An academic need not be “committed” to an institution beyond their contractual notice period. But a postdoc who spends years at an institution that can only offer them dribs and drabs of extension is showing a very strong commitment to that institution indeed.
Myth: postdocs are provided for by their PIs, so do not need representation, recognition or autonomy. There is a widespread strange belief that a postdoc's PI will “speak for them” and reliably look out for their interests. Again, this came up in certain governance debates in Cambridge. It is obviously false; a PI is only ever a partial ally, and can just as easily be an adversary. Yet these debates threw up bogus arguments hilariously reminiscent of those opposed to female suffrage—exclaiming in outrage, “they will just vote the same way as their husband!” and in another breath, equally outraged, “they might vote a different way than their husband!”. (Yes, this was a real argument of the day.)
Myth: increase in postdoc numbers is somehow a “natural” phenomenon. It's easy to encounter the belief that some sort of bumper PhD harvest, no doubt caused by a mixture of climate change and modern agriculture, has led to there being “too many people chasing too few positions”, and that is why so many people are employed on exploitative terms. This is an appealing folk theory, but it simply does not explain what is happening. Positions are not created by nature; they are created by money, spent according to policies. Suppose there are many qualified candidates competing for a fixed number of jobs. Of course, the more candidates there are, the more competition for the jobs. But it doesn't follow that the jobs' terms will become increasingly exploitative, such as being offered in shorter term, with less pay, lower status and fewer benefits. That requires a separate decision to degrade the offering. That is exactly what's happened in early-career research, by turning the knob in favour of creating only these lesser positions. Why so? It's the same story as in the wider economy these past decades: maximising the exploitable workforce, concentrating capital (research funds) among relatively few incumbents. Anyone who tries to explain it purely as supply and demand, or even “nature”, is either ignorant or is trying to sell you something. (Perhaps they're Liz Elvidge, “Head of Postdoc Development” at Imperial College London, who has a book to sell? I haven't read it, but based on her NPM 2017 performance, I assume it's peddling nonsense like this.)
Myth: postdocs are pro-postdoc. Nobody in their right mind actually wants to be a postdoc per se, with all that that implies. People mostly become postdocs because they want to do research. If there were fewer postdoc positions but overall a better path for research careers, few current postdocs would oppose it.
Myth: “nothing can be done”, or “there isn't enough money”. This is the academic version of Theresa May's “magic money tree” line, and is equal nonsense. The issue here is not the amount of money, but about how the money is spent. Policy knobs are very obviously available, but are being turned only in the wrong directions. This is a failure at the top, since that's where the knobs are. All this is outwith the control of research councils, who (despite their many gaffes) just allocate the budget they're given in the ways they know how. The blame lies with central government. In 2002, the House of Commons Science & Technology Committee produced an astonishing report which is entirely damning of the status quo and skewers the problems of short-term research positions. Government's response was a case of fingers-in-ears. Sadly the report dates itself by its naive optimism that the EU Directives I mentioned above would help; we now know that they can be bypassed. In the 17 years since, we've had no action, beyond creation of another worthless pile of paper.
Myth: postdocs just want to stay in the same job or city forever, but that's clearly unsustainable. It's particularly easy to encounter this belief in Cambridge. But the number of postdocs in Cambridge is a function of money, not of wishes. What's really unsustainable is piling all the research money into a small number ever-fatter institutions, on terms that permit only junior and short-term appointments. These institutions gain a large workforce skewed towards the relatively young and exploitable. Later these workers face a vexed choice: either be spat out to make room for the next lot of eager young things, or (if you're lucky) project-hop to another exploitative job in the same city or institution. In contrast with typical PhD-age life stages, postdocs are generally old enough to have put down roots, or to want to. Special as Cambridge is, it is nonsense to credit it with what primarily a desire for stability in one's personal life. Funnelling the bulk of research money to a select few institutions, and primarily on a project basis, is the underlying mistake.
Myth: institutions are doing what they can to support postdocs. In fact the big institutions are heavily invested in suppressing postdocs' career development. Our “leading” universities are the prime foot-draggers and non-movers in this game, and it's not hard to see why: their senior academics profit from cheap highly-skilled labour serving their research empires. Government will only change tack if academics speak to it, but those with the controlling voice have a vested interest. Of course, these already-established research agendas are not necessarily the ones most deserving of support. And even worse, project-based funding bakes in huge inefficiencies which harm outcomes.
Myth: increase in postdoc numbers is essential to creating an agile, global workforce of the future. This sort of neoliberal nonsense is popular among administrators buying the usual wrong assumptions of elasticity and fungibility of people—in short, treating people like a commodity. But on the ground, it's clear that this is a poor model of how research works. Thinly sliced short-term project-based funding not only creates poor-quality jobs, making for unhappy people, but also gives poor research outcomes. Despite the (mythical) “bumper PhD harvest”, (we) academics tend to bemoan how hard it is to find “a good postdoc” to work on their Highly Specific Project X, starting at Highly Specific Start Date D. With those constraints, that's hardly surprising. So begin the compromises. Many postdoc appointments are major compromises on both sides. Sometimes it even works out. But the failure modes are many: people don't fit the project or the PI; they become unhappy; they jump ship or jump career. Then someone new gets hired on an even shorter contract! No doubt the leaving postdoc also spent a good chunk of their work time applying for other stuff. As a result of this churn, much time is lost and much research goes unfinished or unwritten-up; this is “academic backlog” mentioned at length by Dorothy Bishop here and in this talk. Many small grants also push absurdly detailed administrative work onto PIs. All in all, it's an insane way to spend our research budgets.
Given all these problems, what should we be doing? I believe we need a substantial transfer back to core funding and away from research councils. In the UK, a little-known story of the last 40 years has been a series of transfers from core funding to project grants. Where forty years ago it was 60:40 in core funding's favour, now the balance is reversed, and the “match funding” demands of 80% FEC makes the effective core funding level far lower. My investigations suggest that this has not been driven by policy (with one notable exception), so much as it has occured through accidental drift (over roughly three further distinct periods). To fully reverse this, and provide true core funding, we must eliminate the de facto use of core funds as match funding for projects. Projects must be costed at 100% FEC, even if that reduces the fundable volume. In any case, that fundable volume should be reduced! The balance must be made up by increased core funds that institutions can use to hire and retain researchers on better terms, not simply to top up their projects. I'll write more about these issues in a future post.
I fear another “accident” is brewing. Among UK academics disgruntled by REF, it's increasingly popular to say that we should just allocate the whole budget to research councils. No matter how flawed the REF, wholesale transfer to research councils would be a terrible mistake. REF has problems, but the dominance of project grants creates far more problems. It is the over-use of projects, with their thin slicing, and not limited overall spending, that has weakened career structures. In any case, academic time spent writing and reviewing low-success-rate grant proposals dwarfs that spent on REF. The only way to provide sensible careers is an institution-oriented and actively redistributive model making proper use of core funding. It does follow that REF or its successor must leave behind its current rhetoric on “excellence” and emphasis on scores (metrics), since these also exacerbate preferential attachment. Instead it must actively spread out core research funds according to a broad assessment of an institution's capacity for quality research, including for potential and planned growth. Research funding should be a question of steady nurture and wise investment, not a laissez-faire market for powerful rentiers to battle over. The UK is fortunate, in that there is (for now) no shortage of research talent wanting to work here. It will only flourish if we can provide not just jobs, but careers.
[/highered] permanent link contact
