“You’re too stupid to do a PhD. You’re useless. No one here likes you.”

This tale of how supervisory relations can break down irreparably and turn nasty was far from an isolated one among the 106 PhD candidates I recently interviewed about their experiences of switching supervisors.

Doctoral supervision can make or derail a PhD. Good supervision enables . Poor supervision isolates candidates and can .

In a qualitative study supported by the , I interviewed doctoral candidates from 25 Norwegian higher education institutions who had initiated a change of one or more supervisors.

51�Թ�

Their accounts offer a stark insight into what happens when supervision fails and how hard it can be to seek change.

Although the data come from Norway, there is little reason to think these cases are unique. , covering 3,785 candidates in 107 countries, found that almost 40 per cent reported experiences of discrimination or bullying during their doctorate, most often from supervisors or senior staff. Only 28 per cent of those who had experienced bullying or discrimination felt they could raise concerns without risking personal repercussions.

51�Թ�

Survey data tell a similar story about supervision quality. Most PhD candidates are satisfied but between 10 and 25 per cent rate their supervision as unsatisfactory. , 12 per cent report being somewhat or very dissatisfied, yet only about one in 10 of them have actually changed supervisor. A found that nearly a quarter would choose a different supervisor if they could start again.

In my qualitative study, these numbers come to life. Candidates described absenteeism, weak academic support, shifting expectations, hierarchical control and, at the extreme, harassment and abuse. One participant said: “I was afraid of what my main supervisor might do, because they had said they had many contacts in the field and could make my life difficult.” Another recalled: “I was called dyslexic and told I could never finish. I left the meeting in tears.”

Several participants reported explicit threats: project sabotage, loss of authorship, damaged references or blocked access to data. One was ignored for months despite repeated requests for meetings. Another felt compelled to bring witnesses to every interaction. An international candidate was told to speak to no one but their supervisor and warned that they risked being sent home if they did not comply.

The consequences were serious: anxiety, depression, sleep disorders, chronic pain and long-term sick leave. One interviewee told a GP, “If I have to stay here, I might end my life.” Many described dreading emails or corridor encounters, and long periods where “everything stopped” in terms of writing and publishing.

On paper, changing supervisor looks administrative. In practice, it is risky and stigmatised. Candidates fear being labelled “difficult”, losing references or being excluded socially. As one put it: “I had a great network, and then suddenly no one talked to me anymore.”

Procedures were opaque and uneven across institutions. Many candidates carried the full burden of initiating the change, dealing with multiple offices and retelling their story repeatedly. The response of academic leaders varied from quick, supportive action to passive minimisation or overt protection of senior academics.

51�Թ�

Some institutions prioritised reputational risk or promotion cases over PhD candidates’ welfare. One candidate said, “There was a very strong determination from the leadership that my main supervisor should continue no matter what. Because, after all, it’s a stepping stone in academia and part of becoming a professor.” Another was asked to sign a non-disclosure clause not to talk to anyone about what had happened in the project. Some were excluded from research groups or lost authorship rights over their own data. Others were told to behave or leave.

Supervisors can, of course, also experience frustration with candidates and feel poorly supported by their institutions. However, the power imbalance is undeniable. Supervisors often control access to funding, co-authorship, networks and references, sometimes while also holding formal leadership roles. Incentive structures can further distort interests, for instance when supervision counts toward promotion or publication targets.

51�Թ�

In this environment, many PhD candidates felt powerless, isolated, and afraid. Few knew their rights. Some were too scared to contact union representatives. Others were told that switching supervisors would create “too much fuss”. Many therefore stayed silent for too long.

Crucially, most of those who changed supervisors reported improved mental health, renewed motivation and restored progress. As one said, “Everything turned. I got my project back.” Another reflected, “I no longer have to look over my shoulder.”

Supervision breakdowns are not rare exceptions. Doctoral supervision needs to be recognised as complex, high-stakes work and as a core institutional responsibility, not a private arrangement between individuals. The right to change supervisor must be genuine, safe and free from stigma.

Five steps universities should take now: build routine, low-threshold check-ins that surface problems early and include supervision quality in staff reviews; publish clear, visible procedures for changing supervisors, with timelines and responsible roles; ensure access to independent support such as an ombudsperson service; provide mandatory, ongoing supervisor training; and realign incentives so that quality and candidate outcomes, not supervision volume, are what count.

If we are serious about sustainable research environments, we must take better care of the people who do the research. Especially those at the very beginning of their careers.

Johan Kristian Andreasen is associate professor of education at the University of Agder, Norway

51�Թ�