‘Subtle but persistent’ power abuse in academia ‘evades sanction’

Abuses of power in academia are often hard to prove using formal reporting methods but still risk damaging the well-being and professional development of early career researchers, according to a study. 

Universities’ “institutional structures and conditions…enable and disguise power abuse”, says the paper, published in the journal Gender and Justice, which draws on interviews with university staff who have experienced or witnessed gender-based violence (GBV), as well as staff working on institutional responses to such abuse.

“Despite the political and policy significance, the scale, and the social, economic and health costs of [gender-based violence], it remains largely under-reported,” the report notes, referencing a survey from 2022 that found that only 13 per cent of those who had experienced some form of GBV in academia had reported it. 

The paper finds that “abuse [of power] is often subtle but persistent and thus challenging to capture” through traditional incident-based reporting, and that “visible and invisible forms of exclusion and inclusion” shape academia into an “environment that reproduces ideal conditions for power abuse”.

Academia is characterised as being formed of “different forms of dependencies”, including the “asymmetrical relationships” between students and teachers; early career and senior researchers; and the “institutional dependency between the department and so-called ‘star researchers/teachers’, who attract a lot of external funding”.

PhD students and early career academics are identified as particularly vulnerable to such dynamics because they are “studying or working under precarious conditions”.

“Subtle” forms of power abuse include excluding someone from participating in research groups or applications or refusing to provide references, which can have “long-term effects on the individual’s ability to stay in the academic space”. 

It also comes in the form of career advice aimed at “sabotaging PhD candidates’ progress”, and could include negligence by ignoring complaints, piling on workload pressure or withdrawing support, and hindering contacts and collaborations.

The paper describes such incidents as “iterative and hidden”, which do not easily translate into traditional incident-based systems of reporting. 

It says: “While consequences such as withholding information may seem relatively benign, it can be quite a powerful tool for control, especially since networking and scientific collaborations determine early-career researchers’ career trajectories and supervisors often serve as gatekeepers to networks and opportunities.”

“Non-events” can have “severe consequences” for a PhD candidate’s ability to proceed with their work and continue in academia, the study says, adding that it can be “challenging to prove that a lack of response, for instance, is part of a larger pattern of power abuse, that is, part of a system of violence”. 

The researchers, Nicole Ovesen of Uppsala University, Angelica Simonsson of the University of Gothenburg and Vilana Pilinkaitė Sotirovič of the Lithuanian Centre for Social Sciences, recommend that universities should pay greater attention to these subtle and repeated forms of abuse, “instead of waiting for formal complaints of incidents or violence”.

Institutions could achieve this through asking students and staff about experiences of safety, safeguarding and dependencies in student and staff surveys, which “can be a way to capture patterns of abuse which are not incidents-based”.

Universities should also look to prevent power imbalances where employees become “dependent” on relationships with more senior staff on issues such as funding, by ensuring that funding or teaching projects are not tied only to one specific individual. 

“One way to improve the effects of strong ties of dependency for PhD candidates is to ensure that they are able to change supervisor without repercussions and that there are always several supervisors involved,” it says.

juliette.rowsell@timeshighereducation.com