Takeaways from a PhD supervision training

09 Jan 2026

I recently participated in a two-day training on supervising PhD students. It was a good experience, and this post is about some insights I learned/realized.

Training for supervising?

I admit I didn’t high expectations from the training, in the sense that I did not feel I lacked essential skills, or that such skills could be taught in just a few hours. I mostly came because I consider that some people would actually need some kind of training/reminder about how to properly and respectfully supervise students, and it felt hypocritical to have this belief without participating myself. Actually, there was a proposal to make such training mandatory for all PhD supervisors in France, but it was only implemented in specific places like ENS Lyon.

In the end, was it was both enoyable and useful, and I’m glad I decided to participate.

For reference, the training (funded by my employer, CNRS) was “L’impact de votre encadrement sur le projet de thèse et la relation doctorale”, organized by Adoc Metis, and the instructor was Aline Waltzing. All the ideas come from them, I’ve just done a selection and rephrased things in my own words.

Clarity and planning

A recurring theme was the importance of making things explicit. Many aspects of a PhD, such as roles, rules, goals, timelines are often left implicit, and this can create a lot of anxiety.

One of the reasons of this ambiguity is the unique nature of PhD: both a continuation of the academic studies and a job. This hybrid status means that there are no universal standards, unlike for standard students or standard employees. (There was a discussion about terminology: replacing the term “(PhD) student” (or “étudiant” in French) with doctoral researcher (or “doctorant”) to better acknowledge their status as professionals and researchers.)

Since one cannot refer to a standard, it is the role of the PhD supervisor to clarify as many things as possible, starting with practical aspects such as working hours, vacation policies, and expected behavior (e.g. showing up at seminars). Note that these can vary significantly from one group to another.

More generally, it is good to clarify the roles of the student and the advisor, what each of them has to do, can do etc. One aspect of this topic is that in France (but not everywhere), the PhD advisor is not the official manager of the PhD student, only the head of the lab is strictly above in the hierarchy, but in practice the PhD advisor is deciding on the work to be done, and can take action to stop the PhD.

Planning is also a way to clarify On a small scale, scheduling the next meeting at the end of each session gives reassuring milestones. At a larger scale, it is good to talk about the middle and long-term perspectives, and to have some plan. Since the research is mostly about failing, plans will almost always become obsolete, but it is good to have some general direction, and to be able to update it together.

PhD as a learning process

Another important takeaway was to clearly distinguish between the research project and the PhD, as a learning process. The usual way to supervise a PhD is to work on the research project together, and hope that the student will learn by example. But it is good to actually think about the skills that the student will need to learn to become a good researcher, and to consider the research project as a kind of platform to learn these skills.

One issue we discussed was that sometimes the learning process and the project are in conflict. For example, the supervisor would like to develop the writing skills of the student, but because of a looming deadline, supervisor takes over and writes the introduction themselves. More broadly, the advisor may hesitate to delegate advanced tasks to avoid losing time or compromising on the quality (or at least their vision of the quality).

Ideally one would like to start with assuming that the student knows nothing and the advisor providing close guidance, and over time the student gains the skills to take ownership of larger aspects of the project, while the advisor moves to a more consultative role.

Some general classifications

I also found the following classifications insightful and worth sharing:

Psychosocial hazard or work stressor: work intensity and timing of work, emotional labor, lack of autonomy, conflict of values, job insecurity
Sources of motivations: Sense of competence, valuing the activity, sense of autonomy, perception of control, sense of belonging.

Participants

We were around 10 academics at the training, from diverse areas of science. Basically, there were people like me, with little experience, and people who were more experienced but had faced issues recently in supervising. We realize that we are unprepared to supervise students: for most of us, the only PhD we have closely observed was our own PhD, and even with experience, every PhD presents unique challenges. It does not help that most of us are not natural managers: we were selected for our research expertise, and often we are used to working alone.