Wednesday, November 18, 2009


By popular demand (= 3 recent mail requests), my thoughts on co-advising, a topic I have only touched on obliquely before (as far as I can remember, anyway):

One of the obvious benefits for a co-advised student is to have a somewhat high level of interaction with two (or more) professors who can help the student's research, each in a different way. But is that ideal ever attained? How does it work?

My views are mostly from the point of view of the co-adviser rather than as the co-advised. I was briefly co-advised at the beginning of my grad school years, but one professor (let's call him "the sane one") was my main adviser and the other (let's call him "the insane one") was fortunately not so much in control of my destiny. If the insane one had been my sole adviser, I might have quit grad school, or at least left that particular one.

On a few occasions when problems with the insane adviser were particularly severe, I discussed the situation with the sane adviser. He mostly gave me lame advice, but when it really counted (e.g., in an exam), he made sure I was treated fairly.

That's an example in which working with more than one professor can be a somewhat negative experience (it increases your chances of interacting with a difficult person), but as long as one adviser is a reasonable person, you're better off than if you have a single insane adviser.

That's a rather gloomy view of co-advising, so let me hasten to say that as a professor, I have had excellent experiences with co-advising, and I think many of my students have enjoyed being co-advised as well. I co-advised as an assistant professor, but I also made sure to advise some students as sole adviser, because I knew that my department(s) valued this. My co-advising has increased in recent years because I have compatible colleagues with whom I enjoy co-advising.

I think that my co-advising experiences have mostly been successful (says me) in large part because I co-advise with compatible colleagues. I think the experience of being co-advised is enhanced if the co-advisers get along with each other and perhaps even collaborate with each other. This isn't necessary, but it helps create a more interactive research environment for everyone.

For research that is highly interdisciplinary, it can be useful to have multiple advisers in different fields, but if one of your advisers is in another department somewhere else on (or off) campus, it might be a good idea to work out a specific plan for interacting with that adviser; e.g. attend group meetings, take a class, schedule some individual meetings. Also, find out your advisers' research expectations; don't assume that all will have the same ones.

When I co-advise students within my department, both advisers have equal status as advisers. I have, however, co-advised students in other departments with colleagues in those other departments, and in that case the other adviser is the de facto 'main' adviser, although we have equal status on the forms.

There are many possible variations in co-advising relationships, with the main factors being the compatibility of the advisers with each other and with the student (i.e., personality factors), the advising styles and expectations of the advisers, and the student's willingness to take some initiative (but not too much; see below) in communicating with multiple advisers. I think these factors are more important than whether the advisers are in the same or different fields/departments and whether one adviser has more responsibility than another.

The fact that I only co-advise with compatible colleagues doesn't mean that we all have the same approach to advising or that we have the exact same type and level of interaction with our students. In fact, more than one of my co-advised students has said, with respect to a particular colleague with whom I have co-advised, that they wish they could "average" our personalities into one ideal adviser instead of being driven somewhat crazy by our different personalities.

In this case, our students are not saying that one of us is a good adviser and one of us is a bad adviser, but instead that we both have positive and negative advising habits and characteristics and that they wish they could experience mostly the positive aspects and avoid the negative ones in each of us. I sympathize with that, but I can also put a positive spin on it by telling them (and myself) that they are learning important people-interaction skills that might serve them well in their careers.

I think that some of our co-advised students have learned to optimize their interactions with us, going to one or the other depending on their mood/needs. In some cases, our students ask us both the same question and then choose the answer they like better, kind of like asking mom and dad a question and choosing the preferred answer. This is (mostly) fine with me because, despite my differences in personality and advising style compared to my colleague, we are seldom in major disagreement about significant issues related to our students.

At one extreme, students may 'fall through the cracks' between or among advisers. Perhaps each adviser thinks/expects the other(s) to be taking care of their co-advised student, but no one is. Obviously there needs to be good communication among the group, such as might be accomplished during a group meeting of advisers and student to make sure that everyone is in agreement about expectations, priorities, and time lines.

The reason I added "mostly" in an earlier statement is because I recall one student who overdid the ask-both-advisers thing. One of the benefits for a professor of being a co-adviser is that you share the time/work of advising. If a student asks both of you the exact same thing all the time and asks you both to do the same thing so as to choose the preferred result, that is not a good use of our time, especially if we have quite a few advisees.

At some point with this particular student, my colleague and I figured out that he was taking the ask-them-both thing to an extreme. I asked the student to try to reduce redundant effort as much as possible and to use the ask-them-both approach for questions/issues that would benefit from different points of view or for document-editing that really required comments from both of us at the same time. He didn't change anything, so the next time he gave us both something to edit that really only one of us needed to see at the time, my colleague and I sat down together and wrote identical word-for-word edits in exactly the same places with the same pen on each of our copies of the short document. Did the student notice? No, he did not. He was pleased that both his advisers were in such good agreement. He never did stop this habit, but his advisers learned to coordinate with each other so as not to duplicate effort when this was not necessary.

But I digress.

Those students who have previously expressed a wish to average the personalities and advising styles of my colleague and me have said that, once it was over and the degree obtained, they were glad for the experience of working with us both, despite some of the challenges.

There are many different views on co-advising in different academic disciplines and even within different departments of the same academic disciplines. Some may encourage co-advising, some may discourage it. Some may not allow assistant professors to advise a PhD student alone, some may think less of an assistant professor who has not advised a PhD student alone. And so on.

I think co-advising is a good thing, though it needs to be appropriate for a particular student's research and career goals. It's up to professors and students alike to do what they can to make it work well for everyone involved, but when it does work well, I think everyone benefits.

(There will be more on this general topic tomorrow, I think)


Confident Female Scientist said...

I think just having the option can save a student. I'm living the "insane one" with no "sane one" to balance out the disaster. Your post does a great job at modeling an example of what and how co-advising could work in a department with few ass holes and fair amount of common sense. I look forward to joining one such department soon. Not fair...actually my department is great, it's the PI!!!!

Anonymous said...

thanks for this, I'm currently being co-advised in several departments and have been struggling to find a solution to keep the balance. The number of advisors are set to increase over the next few months and maybe decrease again so I pretty much need to know how to coordinate my interaction with the advisors pretty promptly as my program requires that we (post-grad students) were to initiate contact between the departments! Though, reports aside I have been asking both supervisors science related questions via e-mail, as I have no idea who would be better prepared to answer those questions; would that have caused you annoyance?

Anonymous said...

I think this is a fascinating topic. I am from a country where co-advising is mandatory (students must have two or more advisors to guarantee funding for the full PhD period) and have experienced or had exposure to most of the situations you describe.

Actually I have always found the concept of a single advisor odd and never really understood the advantages. Is it really possible that any one (busy) person can meet all the needs of a PhD student? I had four co-advisors for my PhD studies (which is unusually many) and most often this was great - all of them had their strengths and weaknesses, and I could always choose to go to the most appropriate advisor for the question/problem/issue at hand. In fact I don't understand at all the disadvantage of advisors having different weaknesses - it is usually possible to avoid dealing with them by going to a different advisor. All of my advisors were helpful in different situations.

As an assistant professor I find the system useful as well. We are able to advise PhD students without having to guarantee their full salary, and there is no fear of what happens to the student if we don't stay at the same institution for their full PhD period. I think it is better for the students as well - any potential risks of "new" advisors can be tempered by having at least one more senior advisor. Right now I am co-advising four PhD students, which I could never do on my own, but I love the change to interact with each of them.

The only problem I have heard up (second-hand) was when two co-advisors had a falling-out and stopped speaking for a significant part of a student's PhD studies. The student ended up somewhat in the middle because they did not wish to cut or reduce ties to either advisor. I don't think this is common however.

I would be very interested to hear what others see as the benefits of a single advisor, since I have not experienced this system.

JaneB said...

Hi, my experience of co-advising is also from the adviser point of view, and it's less rosy than yours. I've had cases where students made an art form of playing one advisor off against the other, or constantly saying that they were doing what the Other Advisor said, and telling OA they were doing what I said, and actually doing very little (when working in both labs). In some cases it takes a LOT of work to make sure that OA and I don't end up being 'good cop bad cop' (where OA's style is more abrupt than mine). It certainly isn't half the work of a sole advisee, more like three quarters...

Comrade PhysioProf said...

In the biosciences, the true cutting edge of scientific advancement becomes more and more interdisciplinary every day. This means that ambitious and clever grad students are more and more seeking collaborative thesis projects that require "co-mentoring".

Institutional or personal restrictions on this are fucking stupid, and serve to stifle both the development of our best students and the advancement of our fields. Accordingly, I embrace such arrangements.

However, these arrangements--as do all collaborative efforts--implicate the sometimes thorny issue of allocation of scientific credit. While the grad student is, of course, the first author on her publications, how do the co-mentors allocate senior (i.e., last) authorship on those papers? This is a topic worth discussing.

K said...

Thanks for covering this. Nice to get multiple points of view

Female Science Professor said...

Anon 4:33: If you ask both supervisors science-related questions by email, cc them both on the same email so that they know you are asking them both. That would not annoy me at all.

Anonymous said...

for my PhD I had two co-advisors, both with equal weight, both in the same department, of the same rank and status. Ego wars between them led to me being a pawn caught in between them in their political games. They had disputes over money (they were also co-PIs on a large grant that was funding my PhD research), each accused the other of cheating them out of hundreds of thousands of dollars in grant money. It was a nightmare for me. One of them would want me to take the project in this direction, the other would disagree just to be difficult and vice versa. I truly was a pawn in their childish games. It's a wonder I managed to get my thesis done and graduate. Finally after a few years of attempting to co-advise me, they had a falling out and from what I understand never spoke to each other again.

Anonymous said...

There is one clear advantage of co-advising (for a student):
If an adviser dies at least you have a chance to graduate.

Anonymous said...

From an adviser point of view, I am now considering whether or not I should involve a collaborator in co-advising the student I am paying. I have hard time collaborating with nitpicking people and I am afraid the collaborator may delay me and the student in certain situations, i.e. I'd like to make all decisions. And I think I will always be the sane advisor (hehe like to think so). I am a big picture person and I think a lot of people focus too much on details or non essential things that take forever so nothing gets accomplished. So not sure about it. Never co-advised before. I am inclined towards no, unless the collaborator specifically asks.

Anonymous said...

In my co-advised situation the "sane one" advised me on general research questions, field work, and thesis reviews. Unfortunately, the "insane one" controlled graduation, publishing, etc. That situation has not worked well.

Kevin said...

Co-advising is very common in our department, both within the department (I was on the thesis committee for a student who defended this week, though I was not either of her advisers) and between advisers (my most recent PhD student to finish was co-advised by a faculty member in a different department).

A lot of our students end up with two de facto advisers: a wet-lab adviser and a computational adviser. Often the third (or 4th) member of the thesis committee provides yet another viewpoint (a statistician, for example).

We've found (from the faculty side) this to be a very workable approach. Our only constraint is that at least one member of the thesis committee must be in the student's department---it needn't be the primary adviser.

The primary adviser is generally responsible for funding the student. Many of the co-advised students have had fellowships for at least part of their graduate training.

Anonymous said...

My bad experience of being co-advised was one that I"m surprised no one has brought up already. In my experience and observations of others, professors over-use grad students as cheap labor. students are always being told to do things not related to their research and which are part of the advisor's job, under the guise of "giving them experience". Things like writing their grant proposals for them, making their class lecture notes for them, making their conference presentations for them. (but you are not the one giving the presentaion, they are)... Or, the professor wants to generate preiliminary data for future grants (which you will ghost write for them) so guess who has to figure out the experiments and do them and get results but will not be given credit for them? as well as train new students, manage the lab, maintain the lab equipment, redo experiments over and over again simply because the advisor wants the data to look prettier for when they give their invited talks (and you are makingn the presentation for them) even though prettier data is not necessarily better data.

sure I do agree that doing some such work is getting experience which is valuable for your future career, such as grant writing and presentation-making experience, even if you're not getting any credit for it. but at some point one has to question if all this extra work being demanded of students constantly means that they have little time left to do their actual PhD related research. In my experience, the answer to those is usually no. Professors in my department, such as my own advisors, have been known often to withold their signatures from the students' theses after the student had successfully defended, not because the thesis needed revisions but simply because the advisor didn't want the student to leave school so soon and stop doing all that extra work so the professors would keep the students on for longer to use as underpaid lab techs, holding their PhD signatures as ransom.

Thus, to me it stands to reason that having just ONE advisor is already like having two full time jobs - one doing your research for which you hope to get a PhD, the other doing all these other never ending extra tasks that your advisor wants you to do for them as an underpaid employee.

Now imagine having this multiplied by two advisors. Both wanting you to do all these extraneous things for them all the time. this rate if this is multiplied by TWO advisors, you will never graduate. Then they blame you for not being productive enough.

That was what happened to me anyway. In the end I quit in disgust and went to a different department and started over with just one advisor so that I could cope with the workload and graduate within a reasonable time frame.