Universities measure almost everything – except leadership

Research leadership is often measured through proxies, papers, grants, PhD completions. Placing these measures in the context of the wider movement for research assessment reform, Christoph Trattner argues a narrative approach could create a clearer picture of what good academic leadership looks like in practice.

Universities have never been better at counting. Publications, citations, grant income, teaching hours, PhD completions: each has a number, a database and a line in the appraisal form. Leadership is the exception. It is named in every modern assessment framework, and yet when a research leader sits down for a performance conversation, the questions are almost always about their own output. I know this because I have been on both sides of that conversation.

In October 2020 I moved from running a research group to directing a national research centre of around one hundred researchers and professionals across universities and industry. The work changed completely. Success now meant finding funding so that somebody else could have a job, helping a junior researcher turn a half-formed idea into a paper, and spending days on problems that would never appear on my CV. The questions in my appraisals did not change at all: what are you publishing, what are you teaching, which grants have you won?

This is not a Norwegian peculiarity, though Norway is a good place to see it. The Norwegian Career Assessment Matrix (NOR-CAM) , launched by Universities Norway in 2021, lists leadership as one of six competence areas to be recognised alongside research output, teaching and societal impact, and asks for each to be shown through results, documentation and reflection rather than a single number. More than 700 institutions have signed the CoARA agreement on reforming research assessment . What is missing is any settled idea of what evidence of good leadership looks like , and so in practice leadership is read off the same proxies as everything else: the positions someone has held, and the papers and grants they produced while holding them.^[1][2][3]

The incentive trap

The consequence is an incentive trap. If a research leader is rewarded mainly for their own papers and grants, the rational response is to keep maximising individual output – to lead the papers, hold the grants and remain the indispensable centre of every project. Universities then wonder why research groups stay organised around the professor, why the group dissolves when the professor leaves, and why so few postdocs emerge as independent researchers. The system is producing exactly what it measures.

The evidence points the same way. When leading scientists die unexpectedly, their collaborators’ output falls by a lasting five to eight per cent , a measure of how much of a group’s productivity can live inside one person. In Wellcome’s 2020 survey of more than 4,000 researchers, 80% of those managing people said they had the skills, only 48% had been trained, and barely half of those being managed had had an appraisal in the past year. Relatedly in Nature’s 2020 postdoc survey , more than 30% of respondents had already held two or more postdoc positions. Leadership is where these numbers are made.^[4][5][6]

Two questions

Almost six years into the job, I can see what survived of my early work and what only existed while I was in the room. With the centre’s funding period ending in 2028, that is no hypothetical question for me. That experience suggests two questions that would serve better than any output count.

The first: what became possible for other people because you were there? Who finished their PhD, who became independent, who led a paper or a proposal for the first time, which collaboration exists because you built it? This is the leadership equivalent of a publication list, and it is entirely documentable.

The second is harder: what would remain if you disappeared tomorrow? Good leadership creates people, systems and collaborations that function without the leader. A group that stops when its head steps out is not evidence of a strong leader; it is evidence of a dependency. Neither publication metrics nor satisfaction surveys answer this. High output can coexist with a group that has no future, and a contented group can be one that has simply been shielded from difficulty.

What could be documented

Between those two questions sit five things an institution could actually record, in the same results, documentation and reflection format NOR-CAM already uses for research and teaching.

People development: who progressed, and to what.

Delivery: did the group do what it undertook to do with the resources it was given? Organisational health: not a satisfaction score, but the minimum condition of a working environment in which people can raise problems – a floor, not a merit, and one only 37% of researchers in the Wellcome survey felt they had when it came to speaking up about bullying.

Collective capacity: does the group now have skills, infrastructure and partnerships it did not have before? Succession: is there someone who could take over, and have they been given the chance to try?

None of this requires a new metric. It requires asking leaders to document these things and asking evaluators to read them, in the way assessment reform already asks for narrative accounts of research. For fixed-term leadership positions, a mid-term review against these questions would be more useful than an end-of-term report nobody can act on.

It cuts both ways

I should be clear that I am arguing partly in my own interest. I hold a leadership role, and much of what I do is invisible in the framework I am assessed under. But the argument cuts both ways. Taking leadership seriously as academic work means recognising it when it succeeds and holding leaders clearly accountable when it fails: when people do not progress, when the group does not deliver, when nothing is left once the leader moves on. That is a stricter standard than the current one, not a softer one. Nor is it an argument about pay or honours; it is about whether the job is measured at all.

Universities already know how to count researchers. The open question is whether they are willing to find out if their research leaders are doing the job – and to act on the answer.

This article gives the views of the author, not the position of LSE Impact or the London School of Economics. You are agreeing with our comment policy when you leave a comment.^[7]