5.9 Graduate Student Supervision

This is not a note about the supervisor; the graduate student relationship; that is elsewhere. This note is about what a supervisor should perhaps consider teaching when developing a graduate student who is considering a future research/academic career.

It is likely that a supervisor will supervise how they were supervised. Depending on the individual, they might have exposure to one supervisor (Doctoral only, no Masters, one supervisor), or multiples (e.g., potentially co-supervision on both Masters, Doctoral). This can be bad, this can be good. We all know extrapolating off a single data point can lead to interesting results, Type I, Type II errors, perceived cause and effect when there is only fluke of timing involved. It is SO easy and natural to do, the least work; just do to others as others have done onto you. And the circle keeps turning. If you had great supervision, this is not a bad thing. But, how do you know if it is great or not? How do you know what best practice is? How do you know how you should have been supervised? Ever thought about it?

If you are self-aware, have a growth mindset, and is inherently curious, you might have looked around, read some, talked to others. You might reflect on what you see, do not see, what was done, not done, and what the implications are. This is in either case; you had great or, how shall we put it, less than great supervision. Either way, someone self-aware etc., is likely to get a broader baseline for comparative purposes. How much one knows about supervision will impact how they supervise. Simple.

What should a supervisor think about when supervising? What skills and expertise should an professor have? How does one get these skills? A supervisor should know more than one way, and have enough skill and real expertise to be flexible. If the graduate student does not know more about the specific research focus than the supervisor, the supervisor probably failed. This means that the supervisor must feel comfortable supervising things they might not be a full expert in - domain, methods.

Here are some thoughts...

• The list below reflects what we think graduate students need to be taught. The list recognizes that there is opportunity for many discussions with a graduate student during their journey.
• The basics of literature reviews. How to judge quality and impact. What clues to look for in articles; the journal might have a great rep, but that does not mean it is safe to assume that all pubs in the journal are great or are error free. What errors are commonly made, what is credible, what is not. How to pick up citations that appear too good to be true, those which are selective or biased in what they are sharing to make their point (not telling the whole story), those which are likely based on the researcher just looking at the abstract.
• How to review and judge other theses and academic creditials. What makes a theses a good thesis? What are the different types of theses? Not all theses are created equal and are of the same quality. What makes an average thesis. Above average. Noteworthy. Hold-your-nose-worthy. How good is a researcher? Which researchers are to be listened to, respected? Which ones are not worthy of top billing?
• How to think through and design, craft a thesis in conscious dialog. Not just write a chapter, get feedback, revise, but a serious discussion about the design and style of writing, sequence, that makes a thesis friendly to an examiner; what examiners look for. On that note, how to do intel on committee members and examiners and how to deal with this in the writing. Same for preparing for any of the major exams. This is teaching them to fish, versus the fish on a platter. You can supervise either way. Tell them what to do, show them, lead them by the nose, or you can teach them how to be independent, ready for post-degree! We have seen supervisors win awards and praise from students because the supervisor made the degree very efficient for them, told them everything to do, how to do it. Did not allow them to make any mistakes. Wow! What a great supervisor. Nice that they got an award for making the students' lives easy, degrees done in record time. What a sad accomplishment. This is not the approach we recommend. We want the students to know how to develop a research agenda and be capable of doing independent research without having a supervisor on speed dial (old analogy, but still works).

• How to recognize and appreciate the strengths and weaknesses of exploratory, preliminary research vs descriptive, normative, prescriptive, and predictive. How can you figure out how mature a topic is from the literature? If the field is mature and well understood, what should you expect to find in a good literature hunt? If the field is exploratory, possibly pioneering, what should you find (or not find)?
• How to figure out the implicit, unstated assumptions made in papers and research publications, either on purpose or by accident. The student needs to have depth and breadth beyond their specific research topic, focus. They need to have time to explore the ecosystems that their research will touch.
• How to read, analyse at depth, not superficially, so that they can hear and listen, that they can look and see. The numbers and experiments are there, all done with mechanical precision, perfectly executed, but the research or point does not make sense in the first place. Is nonsensical. Is smoke and mirrors when it comes to applied reality or value. Note, the researcher might not be aware of this actually, and someone pointing out the potentially useless or invalid is at risk of flashback and burn. In any event, the supervisor should ensure that the graduate student can make sense of what they are reading, what they are writing, and what they are doing. Not to just go through the motions, cobble together several targeted lit reviews, some hastely thought through hypotheses, some contrived experiments, and then pop it out of the oven as a dissertation with numbers proving that x is significant.
• How to control risk during the research journey. For example, if one has a new method, validate the new method against known benchmarks. If one has a new area of application, use traditional, well understood methods and measures. Do not try to use a new method in a new situation and then claim that the new methods provides significant results. Ah, compared to what? If you have a new idea, you need to know how to take a known baseline comparator and make the one incremental change that allows the research to be considered robust and solid. If you have too many new elements, multiple unknowns, it is hard to make any cause and effect statement. There can be multiple 'new' elements in a thesis activity, but the risk needs to be managed and care made to provide a support for dissemination and scrutiny. Decomposition, step-by-step development are good things. Comparators to known, accepted exemplars or baselines are good things. Graduate students need to understand this. We have seen students and supervisors not address this with disastrous results in publications, publications, and tenure.
• There is also the complementary situation of where the student believes that their work is earth shattering and the greatest thing ever. This might actually be reinforced by some premature recognition and kudos. The supervisor is responsible for the student knowing how good the work is, and not what it might be perceived as being by the student. If the student gets the wrong idea, they might push the 'fantasy' work long beyond its best before date, and end up in the world of excuses why no one will publish the work: 'it is too good, others do not recognize it, do not understand it, others are stupid, others are jealous'. Students need to understand how their work stands, how really innovative it is, and what the real contribute might be. The truth might hurt. But then again, failing to be renewed, failing to be promoted, get tenured will also hurt.
• How to create, plan, think about iterative, incremental, derivative research agendas (the norm), and if the supervisor has a track record, how to deal with disruptive, pioneering, long-haul, high-risk research.
• What to consider in building up a body of work, especially if the work is nonconformist, disruptive. Building up a solid contribution in incremental and derivative research is also a key skill to know.
• Disruptive research - what it is, what it is not, what is incremental and derivative research vs. disruptive
• The student needs to understand the risks and issues behind disruptive, truly innovative research and what this implies for publishing, job opportunities, and such.
• To repeat: if the work is actually disruptive, not just a new idea that incrementally moves the field ahead, the student needs to understand what the journey is going to be like. There are serious risks associated with disruptive research and the student you are guiding into this swamp needs to understand them and accept them. Such research needs to be done with the eyes wide open and risks accepted. Again, we are talking about truly disruptive work, not what is simply claimed to be with flowery adjectives. Research can be the 'first' in a field, but that does not make it innovative with a big I, pioneering with a big P, or disruptive with a big D.

• There is disruptive research that is built upon the existing base, is created from the existing base and is not threatening to anyone. This will be somewhat hard, but it is reasonably easy. Things will be slow, but not too slow.
• There is disruptive research that is somewhat parallel, that does not invalidate the existing body of work, but changes the field. This is slower than the type of research described in the last point. It is higher risk as the existing field needs to be educated and enlightened about the new ideas, methods that complement the field. They will need time to embrace them. This will be tricky work to do and might require a second tier insitution as a 'home'. A top tier school might have problems with the slower journey and risks.
• The third type of disruptive research is the most difficult and is described in the following points. It is when the new research threatens the established base, what others have worked on for years, their thesis work, their status in the research community, what their promotions have been based on. This is the truly disruptive type of research and what normally creates 'new' fields. Think of what happened when the flat earth theory was proven otherwise.
• There will be extreme difficulty with publishing, being recognized. It can take 6-8 years or longer before the work is recognized, before you have a body of work that others take note of. This long-haul research is difficult for most to understand and the normal systems do not support it, enourage it, or recognize it. In fact, you will probably have been fired from the same institution as a junior faculty member that later hired you as a superstar.
• Journals and conferences will not likely have the necessary expertise to deal with the work. It can fall under the 'we have never published anything like this before', 'we do not have referees who can referee it', categories. This might be a sticking point for the top outlets. Lower tiered outlets might be more willing to publish risky, controversial work.
• There are some institutions that will take on high risk performers, but it is also expected that the high risk individual can publish in the traditional outlets, be accepted as member of good standing in the establishment. If the new research throws the old results and methods under the bus, the journey is not easy. If the new research is complementary and allows the old researchers to look good or at least maintain the status quo, there will be peace. If the work is disruptive in the nasty way, it will be important to pick the place to work after graduation with care, and have a strategy for how to introduce the work, get believers, and to slowly proceed. If you go too close to the candle, your wings will catch on fire. You can do great work at a 'lesser' school, and a 'lesser' school is perhaps the best place to start really disruptive research.
• The student will need to know how to create a research portfolio of traditional, derivative work while exploring disruptive and potentially pioneering work. Note, it rarely, ever works to start off saying 'I am going to be a pioneer, and start a field.' Nope. Does not work that way. Questions and ideas will lead the way to research activities and often pioneering work is only seen as pioneering and different in hindsight, or after a critical mass is reached.
• How to be a post-doc.
• How to be a junior faculty member. How to network, and how to find a mentor moving forward.
• What to look for in graduate students, potential collaborators, possible mentors. For students, how to read the tea leaves in transcripts and letters of reference. For others, how to figure out what they do, how they do it, and does it make sense to work with them? Can you work with them? What is in it for each player? How to pick out the key signals, signals of quality, potential.
• How to apply for grants, how to play the game. What is needed for credibility. How to set the hook, imagery, value statements. How to establish your credibilty.
• How to prepare a student for exams, presentations, interviews, conferences, etc. This continues the theme. You, as supervisor, prepare your student, but do you explain your reasoning, discuss alternatives, justify choices, dig deep into the process. This is the thinking you need as a great supervisor and it should be applied to all parts of the student journey.
• How to prepare for promotion and tenure reviews. Again, how to play the game. What common mistakes to avoid. For example, when to get papers into the pipeline so that they might be accepted by the time you need them. What choices to make that contribute to the criteria and what choices do not. When does the criteria change - e.g., teaching school versus research. When is something like a book or book chapter valued, when not. When is a conf presentation of value, when not.
• What activities can be considered to learn more about your field, keep informed, keep current. How to become known in the field, if that is important to you. You can leave the journey to chance, or you can make your own luck, success. For example, what can you do at national or international conferences, work shops to help out. This will help you get noticed. What symposiums, what workshops, where to go and visit; these can help you gain a broader knowledge base and help with your press kit.
• How to plan sabbaticals, what can be learned on sabbaticals. How to do professional development after the PhD.
• How to develop their teaching, how to work with teaching assistants, other course assistants, and of course, how to develop graduate students of their own. How to develop the depth and breadth of knowledge, skill needed in each case.

Students also have responsibility and a role to play in their own learning, but the supervisor has to be aware of what they can learn on their own and what they need help with. For example, how does a student learn how to supervise a student? This cannot be efficiently or effectively learned just by being a student.

You might be a newly minted PhD. You might not be. That is irrelevant. It is about how much you know about supervising, what, when, how, and why. You should think about how much of the above do you know and if it is adequate. Do you know more than 'one thing', do you know different models and comparators. Do you know how to interpret and evaluate different contexts and can you teach the student how to fish? Taking on a graduate student is a big responsibility and it is more than just the research topic, the science, the theory, the hypotheses, lit review, methods, experiments, results, analysis, discussion, future research, and conclusion. It is not just about the exams. Even if not required, seriously consider co-supervision with an experienced pro who can share with you as much of the above as possible. Your students deserve no less. Just because you might be authorized to be a sole supervisor via some set of modules that focus on ethics, conflict resolution etc., that does not mean that you know how to supervise. Think about it. Please.

Further reading

• Donald, J.G., Saroyan, A., and D.B. Denison (1995). Graduate student supervision policies and procedures: A case study of issues and factors affecting graduate study, The Canadian Journal of Higher Education, 25, 3, pp. 72-92.
• Abiddin, N.Z., Hassan, A., and A.R. Ahmad (2009). Research Student Supervision: An Approach to Good Supervisory Practice, The Open Education Journal, 2009, 2, pp. 11-16.
• Azuma, R.T. (1997, 2019). So long, and thanks for the Ph.D.!, 4th revision, UNC. https://www.cs.unc.edu/~azuma/hitch4.html