The Rejection Letter Nobody Reads Twice
You open the portal on a Tuesday morning, coffee going cold, and there it is: three sentences. Two of them are boilerplate you have seen before, on a different rejection, from a different panel. The third is specific enough to sting but vague enough to be useless. Somewhere in the small print beneath the feedback box, a deadline is ticking: fourteen days, or twenty, or thirty, depending on which funding body has just declined your grant. Most applicants never touch that deadline. The window closes. The programme dies. And the scientific record gets shaped not by what was wrong with the work, but by what was wrong with the process for contesting it.
That is the mechanism. It runs underneath the peer-review debates, underneath the impact-factor arguments, underneath every think-piece about how science is broken. The internal grievance structure of a funding body (the specific rules governing who can appeal, on what grounds, within what timeframe, and before which internal body) is one of the most consequential and least examined levers in all of organised research.
The Architecture of a Quietly Closed Door
Most major funding councils operate some version of a two-stage review: an initial panel assessment, then a ranked shortlist that goes to a funding committee which allocates the actual money. A grievance procedure, where one exists at all, typically sits after both stages. That placement matters enormously.
By the time an applicant can formally object, the funding round has often already committed its budget. An upheld grievance rarely means the programme gets funded. It more commonly means the applicant receives a revised written rationale, or a commitment that their next submission will receive closer attention. The remedy is prospective and vague. The harm is immediate and concrete.
The grounds for appeal narrow this further. Most bodies restrict formal challenges to procedural irregularity: a declared conflict of interest that wasn't managed, a reviewer who demonstrably lacked domain expertise, a scoring error that can be shown arithmetically. Substantive scientific disagreement, the applicant who believes the panel simply got the science wrong, is almost universally excluded. Funding bodies defend this on the grounds that second-guessing expert panels would undermine the whole system. The defence is coherent. It is also, in practice, a structural guarantee that panels are insulated from accountability for their judgments.
Consider what that means in a contested or emerging field. A panel stacked with researchers from one methodological tradition will tend to score proposals from rival traditions lower, not necessarily from bad faith but from genuine epistemic disagreement about what constitutes good work. An applicant from the disfavoured camp cannot appeal on those grounds. They can only appeal if someone failed to declare a competing interest, or if the scoring sheet contains an arithmetic mistake.
Two Researchers, One Funding Body, Very Different Outcomes
Take two early-career scientists: call them Dr. Adaeze Okafor and Dr. Martin Lenz, both submitting to the same mid-sized national research council in the same funding round. Okafor's proposal sits in a well-established subfield where three of the five panel members have published extensively. Lenz's sits at the intersection of two disciplines, and the panel has no specialist in either overlap. Both proposals are declined. Both receive three-sentence feedback.
Okafor's feedback, while terse, maps onto the scoring criteria. She can identify what to revise. She resubmits in the next cycle and succeeds.
Lenz's feedback is generic because the panel genuinely didn't know how to evaluate the work. He requests the full reviewer comments under the body's transparency policy. It takes eleven weeks to receive them. By that point, the grievance window has closed. His programme does not get revised and resubmitted; it gets abandoned, because his postdoc contract ends before the next cycle opens.
Nothing illegal happened. No procedure was violated. The grievance structure simply offered no pathway that could have helped him.
What People Actually Get Wrong About Peer Review's Fairness
The standard defence of funding panel decisions is that peer review, for all its flaws, is the least-bad system available. That is probably true. The problem is that the grievance process is sold as a corrective to peer review's failures when it is, structurally, nothing of the sort.
A genuine corrective would need to engage with substantive scientific questions. It would need adjudicators with relevant expertise, independent of the original panel, and remedies that could actually restore a declined programme's chance of funding. Almost no grievance process at any major body meets all three criteria. The processes that exist are designed, quite explicitly, to protect institutional decisions from challenge. Not to catch scientific misjudgments.
This is not a conspiracy. It is a natural consequence of how institutions build appeals processes: lawyers draft them, risk managers approve them, and the goal is procedural defensibility rather than epistemic accuracy. The result looks less like a safety net and more like a moat, elegant from the inside, impassable from the outside. Science administrators are not uniquely cynical; they are doing what every large organisation does when it designs a complaints procedure.
The result, though, is that a funding body's grievance architecture acts as a filter on what science gets a second chance. Programmes in established fields, with legible methodologies and well-networked applicants who know how to frame a procedural objection, survive at higher rates. Programmes at disciplinary edges, from researchers without the institutional knowledge to spot a reviewable irregularity, die quietly. I find this genuinely difficult to defend, and I think anyone who looks at the incentive structure honestly will struggle to defend it too.
The Pressure Point Nobody Is Watching
Reform conversations in science funding tend to cluster around peer review itself: open review, registered reports, portfolio diversification, lottery elements. These are worth having. But they largely leave the grievance architecture untouched.
A few things would change the calculus without requiring wholesale reinvention. Extending appeal windows to match the realistic timeline for receiving full reviewer comments would be one. Allowing substantive appeals to a small, genuinely independent scientific panel, separate from the original committee, would be another. Publishing aggregate data on appeal rates, outcomes, and the disciplinary distribution of declined programmes would at least make the pattern visible.
None of these is technically difficult. Some funding bodies have moved in these directions: the European Research Council has experimented with more detailed feedback requirements, and the Wellcome Trust has published refusal-rate data by career stage. But these are exceptions, and even they stop short of a grievance process with genuine teeth.
So here is the question worth sitting with: if a funding body's leadership genuinely believed its panels were systematically wrong about whole categories of work, would the current appeals structure catch that at all?
It wouldn't. Because most bodies, whatever their strategy documents claim, have built a process designed to catch procedural errors and nothing else. That is a choice, dressed up as neutrality.
The research programmes that get quietly buried are not, on average, the weakest ones submitted. They are the ones that had the worst luck with panel composition and the least access to the procedural knowledge needed to fight back. Science has chosen to tolerate that selection effect, and the grievance form is where that choice becomes concrete.