Balancing prestige and health: why and how to avoid toxic research groups
“At some point, we all have experienced the fact that being lucky is equally as important as being smart and passionate.”1
Since I started my studies, I’ve encountered several victims of abusive supervision. I take this blog post as an opportunity to make people aware of what a toxic research group looks like and what research says about how we can get rid of such Principal Investigators (PIs). This post aims to avoid the bystander phenomenon.2 A bystander is a person who evades taking action against bullying by continuing “the normal state of affairs” and can thus strengthen the bullying act solely by “looking away”.3 To build a healthy and prestigious university, abusive acts have to be seen, discussed, and prevented from happening again. Cutting students’ flow to abusive research groups, providing tools to report and act against such behaviours, and restructuring the academic working environment are the levers we can use towards that goal.4,5
To prevent PhDs and PostDocs from joining toxic research groups, this post gives important advice for what to do before joining a laboratory. Furthermore, the goal is also to warn students about what will happen to them if they join prestigious “superstar” principal investigators who are the subject of backstage abusive stories spread within the scientific community. Lastly, I provide insights into why students remain in such labs and summarize plans proposed in the literature to reform PIs or remove them from academia.4 Information presented in this blog post comes from literature with the intention of keeping a neutral tone.
“Awareness will help students recognize warning signs for themselves and peers, and hopefully counteract the distortions that these behaviors are their fault.” – Cordelia Palitz, clinical psychology doctoral student at The Wright Institute in Berkeley, CA.
For the students/PhDs and PostDocs undergoing abusive events, an anonymous tool to report at ETH is here. Other tools for reporting misconduct at ETH can be found here.
For students who were victims of physical or sexual assault, external counselling can be found here or at the psychological counselling at ETH, here.
If you are already in a toxic lab, insights from Matthias Rillig, professor of ecology at Freie Universität Berlin, about what to do can be found here or from the Nature journal here. Personal histories of other students who went through these events, along with their decisions regarding the situation, are here.
You might think that abusive PIs are stories of the past century, that since then, formal policies have evolved rapidly to provide committees or co-mentors as effective safeguards.6 Unfortunately, literature is filled with recent reports of failing protection systems and of universities protecting abusive PIs.7–10 Why is institutional betrayal happening? The act of silencing harassment emerges from the desire to protect the institution’s reputation and financial interests and/or because of the “old boys’ club” system, where colleagues personally trust the abusing individual and defend him/her.8,11 To protect yourself from abusive PIs and to avoid being a bystander, read more by opening the paragraphs below.
The power of a good mentor.
In most STEM fields, a principal investigator (PI) of a research group or a faculty advisor is responsible for guiding and supervising the work of graduate students during their graduate research, with the implicit expectation that the faculty member serves as a mentor to the graduate student.5 Multiple studies have shown the power of supportive mentorship, demonstrating that mentoring is positively related to graduate students’ persistence in academic research, while also influencing measurable outcomes such as research productivity, number of publications, conference participation, and sustained interest in research.12 In fact, graduate students’ relationship with an advisor, according to some researchers, has the biggest influence on the doctoral student experience.13,14 The PhD experience is also influenced by supportive relationships at departmental and coworker levels, as they reduce emotional exhaustion and therefore decrease intentions to leave academia.15 An effective mentor is someone who provides direction and addresses issues related to projects or group dynamics. They demonstrate trust, flexibility, and strong communication skills. They help establish connections with colleagues, manage and mediate conflicts between team members, and encourage individuals to build their skills and confidence.12 Supporting a diverse research group is essential, as diversity leads to more creativity and better problem-solving.16 Supportive mentoring differs from “laissez-faire” leaders. The latter is characterized by PIs who provide minimal guidance, avoid making decisions, and let employees solve problems on their own. “Laissez-faire” supervision is linked to psychological distress among doctoral students and to bullying perpetration within groups.17,18How is abusive mentoring manifesting itself?
Before any definition, it is important to know that an abusive PI doesn’t need to be prestigious to be toxic. The gender of the PI also doesn’t matter.19 The common trait is the abusive use of their position of power in asymmetry to the trainee’s. In the literature, Tepper defines abusive supervision as: “subordinates’ perception of the extent to which supervisors engage in the sustained display of hostile verbal and non-verbal behaviours, excluding physical contact.” In simpler words, PhD students would perceive mocking, threat, blame, invasion of privacy, or putdowns in front of others from a supervisor as abusive. Improper mentoring also includes PIs promoting a working environment that is pernicious, aggressive, or not conducive to professional growth and productivity.5 Translating these behaviours into acts by the mentor could mean a lack of time and support, insufficient project supervision, oscillating mood or personality changes such as yelling during meetings, thus showing limited interpersonal abilities, and promotion of questionable research practices such as taking credit for others’ work and spreading rumors.20,21 Additional forms of abuse that are not included in the definition of abusive supervision are sexual harassment and assault. These are not covered further in this blog post, even though they were recently reported in Swiss universities by the REFLEKT survey.22 While approximately 10% of working adults report experiencing abusive supervision in non-academic workplaces, surveys in academia report substantially higher rates. Bullying, discrimination, or harassment was reported by 32% of respondents (2518 academic mid-level staff) of the actionuni survey recently run in Switzerland.22,23 These differences are often attributed to structural features of academia, including hierarchical power relations and trainees’ dependence on supervisors for funding, authorship, visas, and career progression. Therefore, academic power abuse has uniquely high stakes.5 Alarmingly, not only is bullying more tolerated in higher educational institutions22,23, but also more bullying seems to take place in prestigious universities.19,21 “QS rankings: ETH Zurich secures 7th place once again” 24Summary of abusive supervision:
Academic bullying is described as a sustained hostile behaviour from one's academic superior including, but not limited to, ridiculing, threatening, blaming, invasion of privacy, putdowns in front of others as well as interference with matriculation and career progress including removing funding, writing falsely negative recommendation letters, taking credit for others’ work and threatening to cancel visa or fellowships.5,19 Abusive supervision, covered in the Employment Act,25 is to be differentiated from sexual harassment and assault. These are forbidden by the Gender Equality Act26 and the Penal Code, respectively.Why do students end up in abusive situations in the first place?
Some students enter unhealthy environments simply because they lacked reliable information at the time of recruitment. Supervisors who later engage in abusive behaviour may appear supportive, motivated, and professional during interviews or short research stays, such as a Master’s project, making it difficult to identify problematic behaviours early on. Once a contract is signed, however, strong power imbalances can make it hard for students to leave or raise concerns. In addition, the abusive relationship is built gradually, where the abnormal gets stepwise normalized, making the recognition of toxic PIs by the students more difficult.27 Other students knowingly join toxic research groups and justify bullying in exchange for prestige.19 They expect reputation spillovers by having the PI's name on publications and CVs. They also hope for faster publication in higher-impact journals, enhanced access to grants, and networking opportunities.5 However, none of these desirable events are guaranteed to happen. To avoid abusive groups, speaking with multiple current and former PhD students before joining a research group is essential. Why ask multiple students? By definition, abusive supervision can be perceived differently by the students. In addition, a toxic PI could act abusively toward certain students only. Finally, abusive environments can encourage competition and discourage openness, meaning that a conversation with a single trainee may not provide an accurate picture since they would want to protect themselves from retaliation or, on the contrary, could aim at gaining their PI’s favours by making a clever/interesting student join the laboratory. Why ask current and former members of the group? Student associations that operate independently from faculty, as well as former group members now working outside academia, can often offer more honest perspectives, as they are less dependent on maintaining professional ties with the supervisor.Summary of possible actions before joining any research laboratory.
This video covers how to spot toxic research groups. The main points are summarized here:- Ask former students: cross-link alumni from the group (usually found on group websites and publications) with LinkedIn profiles. If the target profile is too far from your LinkedIn relations, the connection button is hidden in the three dots next to the follow button. Upon connection, messages can be sent. Otherwise, LinkedIn Premium (trial days) can be used to send a message to those who are not in close LinkedIn relations.
- Explore the lab before applying there, for example, with an internship.
- Be aware of red flags during the interview process. Is the PI interrupting you or dismissive of your questions? Is the PI the only one talking while PhD students can’t say anything? Are they mentioning “working on weekends”?
- Discuss with neighbouring research laboratories about the collaborative environment of your target research group. These students often have no ties with the professor of interest.
- Check the publishing behaviour of the lab. Are there collaborations? Are different students first-authors?
- If you have access to the thesis, check the acknowledgement section. How are the students mentioning their professor and coworkers?
Why are students staying in unhealthy working environments?
Late working hours, weekends at the laboratory, and rare holidays are normalized in academia and can make some students feel more deserving of their doctoral title than others. Negative mentoring experiences are believed to be a necessary or embedded component of science graduate training.5 In other words, students believe that they are truly scientists only if they can survive “trial by fire” and prove themselves to the research community, thus making them accept toxic behaviours more readily.3,28 Other forces apply to international/racialized and minority students for them to stay in toxic prestigious laboratories, since they are more likely to accept and normalize academic bullying.29,30 Under the supervision of a mood-oscillating PI, over time, the person being mistreated may begin to confuse fear with affection, particularly when brief moments of kindness provide a sense of relief. This relief can become a reason to remain in the research group, even when the overall situation is harmful. The dynamic resembles a gambling scenario: repeated losses are endured in anticipation of an occasional win, which feels so rewarding that it sustains continued investment despite the accumulated harm. Within the context of a strong power imbalance, such inconsistent treatment that oscillates between cruelty and care can lead to blurred boundaries and cognitive distortions, with lasting negative effects on self-esteem. Such mechanisms have been extensively observed in intimate partner violence (IPV). However, a recent dissertation by Cordelia Palitz has demonstrated that abusive behaviours enacted by supervisors toward professional female employees are comparable to those observed in IPV.31 To deal with these strong emotions, students try to create a narrative where they retain control.3 They believe they can “deal” with their toxic professor, making their discourse sound like “I can handle it” in other to normalize the suffering. Especially since a PhD stay is defined in time, between 3 to 5 years, students just need to “push through” until they complete their thesis.29,32 Inter-student dynamics also play a role; a strong bond can be created between students going through the same traumatic event. It can feel disorienting to let go of such strong relationships, making it more difficult to take action and leave the abusive environment. Note that the competition in such working groups can also give rise to toxic inter-PhD/PostDoc relationships. Ultimately, staying in a toxic environment feels necessary when no other alternative is seen, or because students feel like they would lose too many years of hard work and feel like losers if they quit. Accepting that one is abused is a difficult step to cross and makes students accept their position longer.What are the consequences of abusive supervision, and what are students willing to risk for prestige under star PIs with unhealthy work environments?
Students are willing to endure toxic workplaces if it means gaining the prestige of working with a famous researcher. This comes with consequences. On a shorter term, the abused target can suffer mental distress in the form of depression, loss of self-confidence, anxiety, and even contemplate suicide. The impacts can manifest as psychosomatic illness such as sleep disruption, concentration difficulties, and increased immune/cardiac risk.33,34 These psychosomatic disorders caused by the toxicity of the workplace also impact work efficiency and laboratory safety.33,35 The Journal of the American Chemical Society acknowledges these effects while keeping the sources of their stories blurred: “We heard stories how students bullied by their peers or supervisors began working at night, alone, to avoid those individuals; […] how stress or pressure to produce results quickly (in both academia and industry) led to improper risk assessment.” 36 Negative experiences with poor leaders have both career and psychosocial effects. Doctoral students whose mentors prevented them from presenting their research at academic conferences indicated that they missed opportunities for career development and networking (career effects) and that this undermined their confidence and identity as a scientist (psychosocial effect).1,4 The mental health impact may persist even years after the traumatic events since: “Targets of abusive supervision report symptomatology that bears striking similarities to those diagnosed with post-traumatic stress disorder.”37 Such discourse is showcased in interviews too.3 In addition, abusive leaders’ behaviours are repeated by followers and are actually more likely to be learned and passed down than positive behaviours.38 This means that new professors who “grew up” with an abusive PI are more likely to believe that their own students must also “pay the dues” of working in the tough field of scientific research.39 And the wheel keeps turning; generations of trauma are passed, while the academic structure lets it happen. Students need to protect themselves since their choice has an impact long after leaving these 4 years of PhD. The PI remains powerful after graduation, for example by bringing criticism to a future employer or diminishing the newly built start-up/research group.40 However, not everything is dark; accounts from the UK have been reported from supervisors who explicitly want to adopt the opposite approach. Such supervisors wish to protect their students from the risk of being traumatised by their superior.41Issues with the academic system and what would need to change.
Bullying in academia is closely tied to how the academic system is structured. Researchers are under constant pressure to prove their value through measurable outputs, such as publications in high-ranking journals.29 This “publish-or-perish” culture increases competition and stress, while strong power imbalances give supervisors significant control over students’ and postdoctoral researchers’ careers. Principal investigators can influence publications, recommendation letters, funding, and, for international researchers, even visa status.4 At the same time, many institutions provide limited support for targets of bullying, and fear of retaliation often discourages reporting, allowing harmful behaviour to continue.19,29 Addressing bullying requires structural change. Universities should:4,12,29,42- publish a transparent report on bullying cases
- clearly define expected standards of supervisory conduct according to the Employment Act
- oblige and cover the costs of professional mentorship training for PIs
- hold faculty accountable by including mentorship quality in promotion, tenure, and salary decisions
- provide formal training for students to raise awareness and intervene
- provide more support to international students, considering their vulnerability to certain bullying (overwork, threats to cancel visa and funding,…)
- support changing students’ advisors to other professors in the same faculty or other universities after reports of bullying cases, no matter the report outcome
- clearly state publicly what kind of support is received when a report is sent
I acknowledge the use of the Literature tool of Edison Scientific to help me cover some of the state-of-the-art research. I read and cross-checked the articles critically before citing them in this post. I continued with my own literature research for in-depth knowledge.
I acknowledge the use of Grammarly and ChatGPT for adding clarity to the text and correcting grammatical mistakes in the final stage of preparing this post. I reviewed the changes critically and revised the writing using my own words and expressions.
Written by Adriana Neves Vieira, proofread by Nina Baumgartner, Natalie Noll, Edith Simmen and Tilly Schaaf, on behalf of WiNS ETH Zurich
