Recently there has been some discussion in the blogosphere about student-advisor relationships in science. Some of this has been triggered by the article by Peter Lawrence in Current Biology (The mismeasurement of science) and some by the recent reports in Nature (here and here) regarding the comparatively low numbers of academic positions relative to the number of new PhDs (and be sure to see also Larry Moran's slightly different and perhaps more realistic take on this).

One of the recurring topics is the pressure that is often put on graduate students by their advisors to publish. Often this seems to be interpreted in a very negative way, with the advisor supposedly viewing students as little more than data-generating and paper-authoring machines (or "indentured labour", according to one of the Nature articles) to be exploited for their own gain. As an advisor, I want to provide some alternative explanations that are not based on such nefarious motives.

At the base of this discussion is the assumption that most advisors actually do encourage/pressure their students to publish -- an assumption with which I will not disagree here. What remains open is the interpretation of why this might occur. There are several possibilities:

1) The advisor does no real lab work himself and brings in graduate students as "cheap, well-trained labour" (as one Nature author put it).

I am sure this happens, especially in some larger groups, but this is not the case in my lab. In fact, one wonders how the students come already well trained. As an advisor, it can eat up an enormous amount of time and energy -- which could easily be spent writing more papers by oneself -- to render graduate students sufficiently competent and confident that one can trust their data as though one had acquired them oneself. Indeed, I resent the insinuation that advisors are self-centred egomaniacs with no regard for student well-being. I take my responsibilities as an advisor very seriously, I care a great deal about the success of my students as individuals, and I work hard to foster an attitude of "my success is your success and vice versa" in the lab. It's possible that the old joke about lawyers ("99% of them give the rest a bad name") applies to advisors, but I do not believe that this is nearly as ubiquitous as it is made out to be.

So, leaving behind the standard interpretation which I strongly reject, let us consider some other possibilities.

2) Advisors need their students to publish so that they can get funds to train more students.

Like it or not, publications coming out of a lab represent a major criterion for whether applications for funding will be supported. No publications, no money. No money, no students. Instead of being a result of megalomania, the pressure put on students to publish can also be the result of a desire to be able to accept eager students into one's lab rather than turning them away. Current students are supported by funds acquired through the labour of others (previous students and/or the advisor), and their efforts indirectly can open possibilities for future students. Think of it as a kind of intergenerational reciprocity.

3) Students want to publish, and the advisor makes sure that they stay on track to accomplish this.

Most of the graduate and undergraduate students with whom I have worked directly have been quite excited by the possibility of seeing their names in print on a high quality piece of work. I would never discourage this, and in fact I do my best to guide them in their research so that in the end it will meet the standards necessary for publication. That may mean extra work well beyond what is required for their course or degree -- but I make a point of screening students as best I can to only accept those who aim higher than the average expectation. It also means that I have to hold the student to a higher standard and to keep the pressure on at times so that their goal of publishing (which of course I share) is achieved.

4) The advisor can easily write single-author papers but wants to write papers with his students.

As with many other advisors, I could easily spend more time working on papers alone or with other PIs. There is something special about writing a paper with a student, however, especially if it is the student's first. It's probably not unlike the excitement of taking a young child to his or her first sporting event, movie, or other activity that a parent has done dozens or hundreds of times and which no longer has that same sparkle of novelty. It is always enjoyable to experience something again for the first time. I remember very well writing my first paper and the excitement of seeing it in print. The only way an advisor can feel that again is to go through it with a new student. In this scenario, students may not quite know how to get a publishable piece of work finished, or may not be thinking that far ahead, but the advisor knows they will be thrilled to have a paper in the end and keeps the pressure on so that it remains a possibility.

5) Publishing will help students who go on in science.

It is a given at this point that having publications is necessary for students to be competitive for future graduate student positions, postdocs, scholarships, fellowships, and eventually jobs. It makes little sense to wait until the end of one stage to publish (e.g., writing up all one's data from one's PhD as a postdoc), and it is far more beneficial to have established at least something of a CV before one starts looking for the next position. Advisors who care about their students' futures will therefore keep the pressure on for them to do high quality work and to put in the effort to publish before they leave the lab.

6) Publishing will help students who do not go on in science.

There is sometimes an implicit assumption that students who plan to go on in science should be treated rather differently from those who do not. Encouraging them to publish commonly falls into this category. Let me point out, however, that not all students know if they want to go on at any given moment (such that these are not discrete categories). This is especially true in systems as in Canada where the MSc and PhD are usually done separately but both involve intensive research projects; some students use the MSc to determine whether they can/want to move on to a PhD. More importantly, it should be obvious that doing work of sufficient quality to warrant publication will help a student no matter what their career ambitions. Why? First, because it shows that while they were in science, they conducted their work at a level high enough to pass peer review and to get into print. Surely a potential employer would recognize this as an indication of intellect, work ethic, and ambition even if lab work, analysis, and writing are not part of the job for which a former student is applying. Second, because inevitably the student will be asking the former advisor for letters of recommendation in the future. If the student has done high caliber work that has been published, the advisor can feel confident recommending her. Whenever possible, I would very much like to be able to write letters about my students like "She went well beyond the normal expectations of her program and completed work of such high quality that it was published as several papers in top journals."

7) If students and advisors are going to invest the time, then the work should be done at a publishable level.

Advisors have very little free time. If they are going to invest it in writing grants to support a student's research and spend the time training students in the methods and analytical approaches of their discipline, going over proposals, attending committee meetings, and generally ensuring that a student has everything he needs to do his research, then there is a reasonable expectation that, in return, the work will be done at a high level. It also makes sense that if the student himself will be investing months or years on a project, that it should be done at a level worthy of publication. Both the advisor and the student win in that case, and the time will have been well invested by both. Making sure that this is true may, of course, involve pressure from the advisor.

(I hope you will forgive a small digression at this point. I don't know how common the misconception is (perhaps it is higher among students?), but advisors are not lazy. Most university faculty work very long hours and are chronically overwhelmed with dozens of duties and commitments. No, they generally do not spend very much time in the lab (anymore), but this is because they are busy writing grants, teaching courses, giving seminars, attending committee meetings, reviewing papers, and juggling countless other tasks -- many of which they consider far less appealing than the lab work that they did as students and postdocs. To have landed one of the scarce faculty jobs, they must not only have done substantial research of their own, they must have risen above the competition in this and other regards. This does not give them permission to view students as data-producing robots (even if they themselves were treated as such), but it does earn them some slack from anybody who might otherwise resent that the boss isn't at the bench very often.)

8) If data are not published, they might as well not exist as far as the pool of human knowledge is concerned.

Data that would otherwise be considered interesting, novel, and important mean nothing if no one knows about them. And if they are never published, then effectively they might as well not exist. The goal of science to expand human knowledge should, in itself, be enough to inspire students to want to publish, but in case that is not enough, there are several practical reasons that students should be expected to generate publication-quality work. One, taxpayers fund most of the research that gets done in academic labs, and they have a right to expect a return on their investment. This does not mean that all science must be done for some specific applied reason, but it does mean that it should not be done solely for the sake of personal interest or in the pursuit of a degree. Two, everyone benefits from scientific knowledge, but unlike most individuals, students have an opportunity to add to it as well -- so long as they make their results widely available. As one professor I know puts it with tongue partly in cheek, "unless you actually contribute something to human understanding of the world, you are a parasite on those who do". Three, not publishing means that other students may waste their time in other labs, trying to develop the same methodologies and making the same mistakes, because they were unaware of the work that had been accomplished already. It is also the case that other students may replicate work that has already been completed rather than expanding on it or focusing on some other issue in need of study. Not publishing essentially means that new knowledge produced by students is lost, and the advisor has a responsibility to prevent this if at all possible.

The point here is that there are many positive reasons why students should aim to publish and why this should be encouraged and expected by their advisors. Implicit assumptions that advisors have only their own selfish interests at heart can do little more than to discourage students from trying their hand at research and to offend advisors who care sincerely about their students. The publication of exceptional research is in the best interest of both the student and the advisor, but this may not always happen without some encouragement and pressure at the right times.