Attention conservation notice: This is a long post. Sorry, I thought it worked best that way. tl;dr version:
- Collaboration per se isn’t undervalued in academic science.
- We should, and mostly do, value good science, rather than valuing the way in which that good science was produced (e.g., collaboratively vs. not).
- Academic science necessarily is “competitive”. But “competitive” in the relevant sense is not the opposite of “collaborative”.
- Different strokes for different folks. People who work in a different way than you (e.g., collaboratively vs. not) are not doing it wrong. Don’t attack them.
I’ve read various pieces and had various exchanges of comments over the years, to do with how collaborative work is valued and how those who do it are evaluated in the “competitive” world of academia. I’ve found some of these pieces and exchanges rather puzzling, as they seem to question or misunderstand points that I took for granted. But maybe I’m misreading what folks are trying to say when they complain that collaboration is systematically undervalued, or that “competition” in academia selects against collaborative work? Anyway, here are my thoughts on that.
Collaborative work isn’t undervalued. The mean number of authors per paper has been rising for many years (see here, here, and here). That’s in part because of the rise in papers with dozens or even hundreds of authors, so it doesn’t just reflect growth in average lab size or changing authorship criteria. Sole-authored papers are increasingly rare, even among non-review papers. A commonly-used index of author “impact”, the h index, is calculated by allowing each author of a multi-authored paper to count the paper as one of their own, which if anything seems like a bias in favor of collaborative work. (note that I don’t actually like the h index; I’m just noting a feature of how it’s calculated) Closer to home in ecology, NCEAS has been hugely successful and inspired a bunch of imitators. Or think of NutNet and its growing number of imitators. And some broadly-collaborative activities, like data sharing, are so valued that they’re now required by many journals and funding agencies. Many funding agencies also have programs specifically aimed at collaborative work (e.g., NSERC’s CREATE and SNG programs). Yes, anecdotally, I’m sure there are people who don’t like collaborative work, or at least certain sorts of collaborative work. But anecdotally, there are people who don’t like puppies. They’re the exception, not the rule. So when somebody writes that in academic science there’s too much competition, and that this discourages collaborative work “such as the Human Genome Project”, I do a double-take. Because, you know, we had the Human Genome Project (and many other massively-influential, widely-copied collaborative projects). I don’t see how all that would be possible if collaborative work was systematically and seriously undervalued.
Universities hire and evaluate individuals, that’s as it should be, and there’s no changing it. Ok, this one almost certainly goes without saying. But on the principle that if something goes without saying it’s best to say it anyway just to be on the safe side, I thought I’d say it. I appreciate that being evaluated relative to others, and having your science evaluated relative to other science, can be deeply unpleasant and stressful, even if it’s done well and especially if it’s done badly. But professional science is not, and cannot be, a race in which there are no winners and everyone just gets a medal for participating.
Let me also say that, in my experience, research universities are pretty good at this (I know little of other sorts of institutions, hence the limited scope of my comments here). See here for a bit of relevant data. In particular, research universities are not opposed to collaborative work in my experience. They want to hire and promote good independent scientists, which emphatically is not the same thing as wanting to hire and promote scientists who work alone. For instance, someone like Brian, who produces a lot of really good science by being very active in working groups—leading some, making substantial contributions to others—is going to be well regarded by any sensible research university (as evidenced by the fact that three different research universities have hired Brian). As another example, Meg and Spencer Hall have a long-term collaboration—which their employers are fine with (as far as I know!), because it’s a productive collaboration to which they both contribute a lot. Neither one is just riding the other’s coattails. That’s what universities don’t want—people who just ride somebody else’s coattails rather than making their own substantive contributions.* Now, one way to demonstrate that you’re not riding anyone’s coattails is to do some solo work–but it’s far from the only way.
I’d say more or less the same for other institutions that evaluate scientists and their science relative to one another, like granting bodies.
A world in which everyone collaborates and collaborative work is highly valued is still a competitive world. My point here is more or less the same as Darwin’s in the Origin of Species, when he emphasized that he meant “competition” and “struggle for existence” in a broad sense. “Competition” or “struggle” for Darwin doesn’t just happen between predators and prey, or males fighting over females, or etc. It also happens when, say, different varieties of flower compete to attract insect pollinators by offering greater nectar rewards. That is, they compete to be the best mutualist, and the best mutualist might well exclude inferior mutualists.
Analogously, as long as there are more applicants for academic jobs and grants than there are academic jobs and grants, as long as universities and funding agencies seek to hire and fund the best people and the best science (by any criteria), and as long as scientists pick and choose what work to pay attention to (again, by any criteria), there’s going to be “competition” in the relevant sense. You can compete by being a “mutualist” and doing lots of collaborative work, but you’re still competing. Whether you like it or not, and even whether you’re aware of it or not. As with the previous point, I don’t know that anyone’s ever explicitly suggested otherwise. But I’ve seen pieces that almost imply otherwise, e.g., many pieces contrasting “competition” with “collaboration”. There’s no contrast, not when “competition” is understood in the relevant sense. Tim Taylor makes this point nicely in the context of economics. Or think of how different collaborative groups raced one another to sequence the human genome, or to discover the Higgs boson. Closer to home in ecology, Carl Boettiger has a very nice comment on an old post of ours, envisioning a world in which scientists compete to produce the best “shared” contributions to science—the best shared datasets, the best MOOCs, the best R packages, etc.
It is not the job of universities to either favor or disfavor collaborative work over other sorts of work. It’s their job to favor good work. You have academic freedom as a university researcher, and that’s absolutely precious. Insofar as universities tell you to work in one way rather than another, they are infringing your academic freedom. Universities should (and in my experience, mostly do) care about outputs, not inputs–your scholarship, not how you did it. It’s the scholarship that has value, so that’s what should be valued. Insofar as universities don’t do that, and devalue certain work because of the way it was done, the solution is not to have them value that way of doing things. The solution is to get them to quit valuing work based on how it was done, and instead do what they ought to do—evaluate the work itself. In particular, if universities undervalue good collaborative work just because it was collaborative, the solution is not to have them value collaboration for its own sake. That would amount to having them overvalue bad collaborative work just because it was collaborative.
For instance, as noted above, Brian’s a self-confessed working group junkie. He should be, and has been, rewarded for that. Not because working groups are valuable for their own sake, but because the working groups Brian’s led and participated in have been really productive and that’s thanks in substantial part to Brian’s contributions. Brian, as an individual, is really good at collaborating. He’s been rewarded for that by being rewarded for the good science he’s produced. In contrast, I’m much less good at collaborating, at least in working groups. I’ve only ever been involved in a couple of completed working groups, one of which I led, and both of which were only moderately successful (due in large part to my shaky leadership in the one case). I shouldn’t be, and haven’t been, dinged for not being involved in more working groups. But I haven’t produced as much good science as I would have if I was better at spotting and pursuing opportunities to do good working group-based science. And I quite rightly haven’t been evaluated as well as I would’ve been had I produced more good science.
Even if we wanted universities to favor certain ways of working over others, we don’t know what ways of working are best, or even if there is a single “best” way of working. And “best” for what or whom? Terry McGlynn wrote a very nice post recently about “academic Moneyball”. You should click through and read it, but here’s the short version. Terry’s concerned that some ways of working, and some skills and attributes of academics, are undervalued assets. In the same way that Major League baseball teams once undervalued the ability of hitters to draw walks**, a real skill that demonstrably helps win baseball games. I think Terry’s got a point, but it’s not quite the point he thinks he has.
The problem with his analogy is that, in major league baseball, we know what team every player plays for. We know more or less exactly what each team is trying to achieve—win as many games this season as possible (it’s a bit more complicated because a team might, say, sacrifice wins now to increase wins in future). And we can estimate, in an objective and fairly rigorous way, how much a skill like drawing walks contributes to winning games, thereby letting us identify objectively-undervalued skills.
In contrast, it’s not so clear what “team” an individual scientist plays for. The department? The faculty? The university? Science as a whole? And don’t say “all of the above” because what’s good for one of those teams could well be bad for another. Just as in major league baseball, taking walks is probably good for your team but bad for professional baseball as a whole (because it’s boring to watch and so drives away casual fans).
Nor is it clear what the goal of a scientist’s “team” is, or even if there is one goal. Different teams can have different goals, and one team can have several goals. For instance, Terry says that both mentoring and teaching are undervalued—but by whom? Different universities value different things to different degrees. I’d argue that that diversity of institutional values is a good thing for academia as a whole. And any given university ordinarily has several incommensurable things it values. That’s why job expectations at many universities are stated in terms of three goals (research, teaching, and service), often with the explicit proviso that you have to meet some minimum standard in all three to be tenured or promoted.
Nor is it clear how to estimate the contribution of a given scientific skill to achievement of any given goal of any given “team”. In particular, there are scientists who’ve made massive contributions to science by working solo, and others in the same field who’ve made massive contributions by working collaboratively. Andrew Wiles proved Fermat’s Last Theorem pretty much on his own, but Paul Erdös was the most famously-collaborative mathematician ever. Closer to home in ecology, The Silwood Circle notes that the ecologists at Silwood Park had massive, positive influence on ecology, and shared certain traits and ways of working. Including oft-criticized traits like competitiveness, ambition, and greater-than-usual willingness to criticize the work of others. Further, they had massive positive influence on ecology because of those traits and ways of working, not despite them. They also exhibited some widely-praised traits, like “seeking out opportunities for productive collaboration”. Again, “collaborative” and “competitive” are not antonyms.
And even within any given department—or any given scientific collaboration!—there will ordinarily be different people who contribute in different ways to the success of the whole. Indeed, many collectives are successful in large part because of the diverse skills and attributes of the individuals comprising them. Division of labor and comparative advantage are really useful! Meg has a good post on this.
That’s not to say that I don’t think scientists or their work can be judged—as I said above, they can be, should be, and will be! But it’s important to recognize that those judgments are just that—judgments. They’re not purely subjective or arbitrary, like your choice of favorite color. But nor are they as objective as judgments about the value of drawing walks in major league baseball. It’s fine for people to argue that some “team”—their own department or university, all universities, granting agencies, science as a whole, whatever—ought to value something differently than it does (e.g., this old post of Brian’s). But let’s admit that that’s a matter of professional judgment, not something on which there’s some discoverable objective truth of the matter.
Working solo does not make you “selfish” any more than working collaboratively makes you a “parasite”. I wish I didn’t have to make this point. But I have seen misguided critics of certain collaborative ways of working call the people who work in those ways “parasites”. And I’ve seen misguided advocates of collaborative ways of working call people who work in other ways “selfish”. They’re both out of line. For instance, I have numerous sole-authored papers on my cv. They’re not there because I’m selfish. They’re there because I had some good scientific ideas that I happened to be perfectly capable of pursuing on my own. The same is true for pretty much everyone else who sole-authors papers. And even if you mean that kind of language in some technical or metaphorical sense, like “selfish gene” is a technical metaphor, you really should avoid it. People will quite naturally take it as a personal attack. I’ve been guilty of this myself, in the context of some of my writing on the tragedy of the peer review commons. I’m trying not to make that mistake again. Fortunately, comments like these are rare in my admittedly-anecdotal experience.
*Not that our evaluation methods are perfect, obviously. For instance, I’m a little surprised that Meg was advised to do some solo work so as to help demonstrate she’s not just riding Spencer’s coattails. I’d have thought it would be obvious to anyone who knows Meg and Spencer and their work that neither is riding the other’s coattails. But I guess not. On the other hand, I’ve certainly seen collaborative papers with scads of authors where I have no clue how to tell if author X made some substantial contribution or was just along for the ride. And so I can imagine situations in which it might be difficult to evaluate an individual scientist and their contributions.
**Don’t bother clicking this link unless you’re a serious baseball fan. And go Red Sox!