How to Write a Grant Proposal

Writing a competitive research-grant proposal is a distinct skill from knowing what a proposal contains. This guide covers the writing process itself — how to build a persuasive narrative around your specific aims, how to write to the criteria a reviewer will actually score you against, and the mistakes that sink otherwise-fundable science. For the full inventory of what a proposal package needs to include (cover page, biosketches, letters of support, and the rest), see CASRAI’s Components of a Grant Proposal; for the step-by-step submission workflow (finding an opportunity, institutional routing, portal submission), see How to Apply for a Research Grant. This page sits above both: it is the foundational entry point for the writing task itself, and the more specific guides linked throughout cover individual components and mechanisms in depth.

Start from the review criteria, not the components list

A proposal is not an essay; it is a document a committee will score, section by section, against published criteria — and the single most common structural mistake is writing to a generic sense of what a proposal ‘should’ look like instead of to the criteria the reviewers on that specific study section or panel are handed. Two conventions dominate US research funding and are worth internalizing before drafting a word:

NIH study sections score applications against five core review criteria: Significance, Investigators, Innovation, Approach, and Environment, each given its own numeric score that rolls up into an overall impact score. See CASRAI’s NIH Study Section entry and NIH’s Simplified Review Framework guide for how that scoring actually works.
NSF reviews every proposal against exactly two merit-review criteria: Intellectual Merit (the potential to advance knowledge) and Broader Impacts (the potential to benefit society), both defined in NSF’s Proposal & Award Policies & Procedures Guide (PAPPG). A Project Description that doesn’t explicitly and separately address Broader Impacts, not just Intellectual Merit, is a routinely cited reviewer complaint.

Foundations and international funders publish their own criteria, sometimes less formally, but the discipline is the same: find the actual scored criteria for the specific opportunity before drafting, and organize the narrative around them directly rather than around whatever structure feels natural to the writer. See CASRAI’s NIH Grants: An Overview and NSF Grants: An Overview for how each agency’s process fits together, and Types of Grants for how funding mechanisms differ more broadly.

The core narrative components, briefly

Every research-grant narrative builds around some version of the following. Exact names, order, and page limits vary by funder — always check the specific funding opportunity announcement over any general guidance, including this one — and CASRAI’s Components of a Grant Proposal guide covers each of these, plus the full supporting-document list (biosketches, current-and-pending-support disclosures, letters of support, data management plans), in structural detail.

Specific aims or objectives — a short, tightly written statement of exactly what the project will accomplish, usually two to three discrete, measurable aims (NIH gives this its own one-page attachment for an R01). Write this first and revise it last: it is the lens reviewers read everything else through, and every other section should visibly serve one of the stated aims.
Significance — why the problem matters and what changes if the work succeeds, argued independently of whether the approach is sound. A significance section that only restates the aims, without establishing the size of the gap or the cost of leaving it unaddressed, is a frequent reviewer criticism.
Innovation — what is novel about the concept, methodology, or technology relative to the current state of the field. Scored as its own NIH criterion; expected implicitly even where it isn’t a separate scored section.
Approach — research design, methods, analysis plan, and rigor — typically the longest section and the one reviewers scrutinize hardest. Each aim should have a corresponding approach subsection, including how you will know if it fails and what the alternative is; NIH’s Rigor and Reproducibility expectations (see CASRAI’s NIH Rigor and Reproducibility policy entry) apply here specifically.
Budget and budget justification — written jointly with the institution’s sponsored programs office, but the investigator is responsible for making sure the science and the budget agree: a budget that doesn’t visibly support the proposed approach (too little personnel effort for the described workload, equipment not tied to a named aim) reads as a mismatch to reviewers even when the science itself is sound. See CASRAI’s Budget Justification Narrative guide and Budget Justification: A Worked Example.

Common reviewer pitfalls

These recur across NIH study sections, NSF panels, and foundation review committees, in one form or another, more often than any single technical flaw in the science:

Aims that aren’t independent. If Aim 2 can’t proceed unless Aim 1 succeeds exactly as predicted, reviewers read the whole project as fragile — one failed experiment sinks the entire proposal. Aims should be able to stand largely on their own.
Approach that doesn’t map cleanly back to the aims. Reviewers actively check whether every aim has a corresponding, adequately detailed approach subsection, and whether the approach section drifts into work that serves no stated aim.
Overclaiming from preliminary data, or offering none where it’s expected. Preliminary data should demonstrate feasibility, not pre-answer the research question — a proposal that reads as already having the results asks reviewers why it needs funding. Conversely, for mechanisms that expect preliminary data (most R01s, unlike exploratory R21/R03 mechanisms), its absence is a common weakness cited in summary statements.
A significance section written for insiders only. Study sections and panels include reviewers outside the applicant’s immediate subfield; a significance argument that assumes deep familiarity with a narrow literature underperforms one that establishes the stakes in terms any qualified scientist can follow.
Ignoring the funder’s own scored criteria. An NSF proposal with a thin, bolted-on Broader Impacts paragraph, or an NIH proposal that never explicitly addresses rigor of prior research and scientific premise, gets marked down on a criterion the writer effectively skipped.
Budget and narrative that don’t agree. See above — this is as much a writing problem as a budgeting one, since it’s the investigator’s job to make sure the two tell the same story.
Missing or generic letters of support. A form letter from a collaborator that doesn’t specify what they will actually contribute reads as weaker than no letter at all in some reviewers’ eyes. See CASRAI’s Letter of Support for a Grant Proposal guide.
Formatting and technical noncompliance. Electronic submission systems reject applications outright for missing attachments, wrong biosketch format, or exceeded page limits — independent of the science. This is a preventable, non-scientific reason for rejection that costs an entire submission cycle when caught late.
Drafting against the deadline instead of the review cycle. A proposal finished the night before submission has had no time for a colleague’s read, an internal mock review, or a sponsored-programs compliance check — all of which routinely catch the pitfalls above before a real reviewer does.

A practical writing sequence

There is no single required order, but a sequence that works well in practice:

Write the specific aims page first, even in rough form, before anything else — it forces the project into a small number of concrete, testable claims and becomes the outline for every other section.
Draft significance and approach in parallel, checking constantly that each approach subsection visibly serves a stated aim and that significance sets up a gap the approach then closes.
Loop in the sponsored programs office early for the budget, not at the end — budget development takes longer than most investigators expect, and a budget built after the narrative is finished is more likely to disagree with it.
Request letters of support and compliance approvals (IRB, IACUC) well before the deadline — these depend on other people’s schedules, not just the writer’s.
Get a critical read from someone who is not a co-investigator — ideally someone who has served on a study section or review panel for the relevant funder, or an institutional grants-development office if one exists. A reviewer’s-eye read at the draft stage catches the pitfalls above far more cheaply than a rejected application.
Revise the specific aims page last, so it accurately reflects what the finished proposal actually argues, not what you intended to argue when you started.

If the proposal is not funded on first submission, most major funders allow a resubmission; NIH, for example, permits a single resubmission (A1) of an original application, submitted after the summary statement from the prior review is issued. Read the summary statement or reviewer feedback carefully — it is a direct, specific list of the pitfalls above as they applied to your actual application, and a resubmission is judged on how directly it responds to that feedback.

Where to go next

This guide covers the writing task in general. For depth on a specific component or funding mechanism, CASRAI’s grants-management content includes:

Components of a Grant Proposal — the full structural breakdown of every required section and attachment.
How to Apply for a Research Grant — the submission process and timeline, from opportunity to portal.
How to Write a Timeline for a Research Grant Proposal — building the project timeline component specifically.
Budget Justification Narrative for a Grant Proposal and Budget Justification: A Worked Example.
Letter of Support for a Grant Proposal.
The Grant Proposal Approval Process — institutional routing and sign-off before submission.
NIH Grants: An Overview and NSF Grants: An Overview.
Data Management Plan (DMP), Biosketch, and Current and Pending Support — the supporting documents most proposals also require.
Grants Management pillar page — the full cluster overview.

Frequently asked questions

What is grant writing?

Grant writing is the process of producing a research-funding application’s narrative components — typically the specific aims, significance, innovation, and approach sections — so that they argue persuasively for the project’s importance and feasibility while explicitly addressing the funder’s own scored review criteria. It is distinct from, though closely coordinated with, the budget and administrative components of a proposal, which are usually developed jointly with an institution’s sponsored programs office.

How long does it take to write a grant proposal?

Timelines vary widely by mechanism and by how much of the underlying research design is already settled, but institutional research-development offices commonly advise starting a full NIH R01-scale proposal at least two to three months before the deadline, with earlier lead time needed for compliance approvals (IRB/IACUC), letters of support, and budget development, all of which depend on other people’s availability rather than just the writer’s own schedule.

What makes a grant proposal competitive?

Beyond the underlying quality of the science, competitive proposals share a few writing-level traits: aims that are specific, measurable, and largely independent of one another; a significance section that establishes real stakes for a non-specialist reviewer; an approach that visibly maps back to each aim, addressing feasibility and rigor directly; and a narrative that explicitly speaks to the funder’s own scored review criteria rather than a generic sense of what a proposal should contain.

Should I write the specific aims page or the full narrative first?

Most experienced grant writers draft the specific aims page first, even roughly, because it forces the project into a small number of concrete, testable claims that then structure everything else — and revise it last, once the full narrative shows exactly what the proposal actually argues.

What happens if a proposal isn’t funded on first submission?

Most major funders allow at least one resubmission. NIH, for example, permits a single resubmission (A1) of an original application, submitted after the summary statement from the prior peer review has been issued; the resubmission is expected to respond directly and specifically to the prior reviewers’ comments, not simply restate the original application.