# How to write a grant proposal, from the funders' own guidance

**How do I write a research grant proposal that meets the funder's rules and review criteria?** To write a grant proposal, answer the funder's own questions in its order and within its limits: a summary, aims, background, methods, expected impact, a data plan and a justified budget. Limits differ widely: NSF allows 15 pages for the project description, ESRC 2,500 words for Approach, SSHRC six pages. Reviewers score against published criteria, so write to them.

Published 2026-09-26 by EdCitation. https://edcitation.com/newsletter/how-to-write-a-grant-proposal

This guide on how to write a grant proposal is published by EdCitation. It is for a research grant application to a funder, not the proposal a student writes before a dissertation, which has its own guide: [how to write a research proposal](https://edcitation.com/newsletter/how-to-write-a-research-proposal). On 26 September 2026 we read the guidance of seven funders in five countries: the NIH and NSF in the United States, the ESRC in the United Kingdom, the European Research Council, the Australian Research Council, Canada's SSHRC, and the Wellcome Trust.

They agree on what a proposal must show and disagree on almost everything else: how long it may be, what counts toward the limit, and how it is scored. The call you answer is the only guidance that binds you. EdCitation's free [Check your paper](https://edcitation.com/check) reads a funder's call into a checklist of limits and required sections, each shown beside the sentence it came from. EdCitation never writes any part of a proposal, and nothing here, or in any tool, can promise funding.

## How to write a grant proposal, step by step

A grant proposal is written backwards from the funder's rules: read the call and the review criteria first, then write each section to them.

1. **Read the call and its criteria.** Note every page or word limit, required heading, deadline and scoring weight. [Check your paper](https://edcitation.com/check) lists them free, with the source sentence; check its list against the call yourself.
2. **Fix the aims.** A few aims, each answerable within the grant and able to succeed alone; see [how to write research questions and aims](https://edcitation.com/newsletter/how-to-write-research-questions-and-aims).
3. **Gather the background.** Find the studies each claim needs with [Find sources](https://edcitation.com/), searching by the claim itself.
4. **Write the methods and plan.** Show that the work is feasible in the time and budget asked for.
5. **Write impact, data and ethics sections** to the funder's own questions.
6. **Cost it and justify every line.**
7. **Build and check the references.** [Cite a source](https://edcitation.com/cite/apa) builds each entry from its DOI; [Verify references](https://edcitation.com/verify-references) checks the finished list.
8. **Write the summary last**, then ask colleagues to read the whole proposal as reviewers would.

## What sections does a grant proposal have?

The structure of a grant proposal is nearly the same at every funder, under different names: a summary, aims, background, methods, expected impact, a data plan, and a budget with its justification. What changes is the order, the heading and the limit.

### Summary and aims

The summary is read first and often by non-specialists. NSF requires a one-page Project Summary in three parts: an overview, a statement on intellectual merit and a statement on broader impacts (National Science Foundation, 2024a). ESRC's 550-word summary is written "in plain English" and is used to find the right expert reviewers (Economic and Social Research Council, 2026). Wellcome's research summary is 200 words and is published unedited if the award is made (Wellcome, 2026). For the NIH, the Specific Aims page stands in front of the science; Fahrenholz et al. (2023) describe it as a single page added to the 12-page Research Strategy of an R01.

### Background and methods

The background shows the gap; the methods show you can close it. SSHRC asks for a detailed description under three headings, Objectives, Context (including the literature review and theoretical approach) and Methodology, in six pages at most (Social Sciences and Humanities Research Council, 2026b). In the form for its 2026 calls, the ERC splits the case in two: Part I sets out the state of knowledge, the questions and the overall strategy in five pages, and Part II gives the methodology, work plan and risks in seven (European Research Council, 2025a).

### Impact, data and budget

These sections are scored, not decorative. NSF's Data Management and Sharing Plan may run to two pages and is reviewed under intellectual merit, broader impacts or both; its budget justification may run to five (National Science Foundation, 2024a). ESRC gives 500 words to data management and sharing and 1,000 to resources and cost justification. Wellcome asks for an outputs management plan of up to 500 words.

## How long can each section be?

Each funder sets its own limits, in pages, words or characters, and they are not interchangeable. The table gives the main narrative, the summary, the rule for references and what reviewers score, as each funder states it.

| Funder and scheme | Main narrative | Summary | References | Scored on |
| --- | --- | --- | --- | --- |
| NIH R01 | Research Strategy 12 pages, Specific Aims 1 (as Fahrenholz et al. report) | Not confirmed at source | Not confirmed at source | Importance, Rigor and Feasibility (1 to 9); Expertise and Resources |
| NSF (PAPPG 24-1) | Project Description 15 pages | 1 page, three parts | No page limit; every author, in order | Intellectual Merit, Broader Impacts |
| ESRC research grants | Vision 500 words, Approach 2,500 | 550 words | Own question, 1,000 words | Vision, approach, data, team, ethics, resources |
| ERC Starting Grant (2026 form) | Part I 5 pages, Part II 7 | 2,000 characters | Outside the page limits | Excellence, the sole criterion |
| ARC Discovery Projects 2027 | Not confirmed at source | Not confirmed at source | Preprints marked, with DOI or URL | Quality and innovation 45%, investigators 30%, benefit 15%, feasibility 10% |
| SSHRC Insight Grants 2026 | Detailed description 6 pages | 1 page | Up to 10 pages, any style | Challenge 40%, Feasibility 20%, Capability 40% |
| Wellcome Discovery Awards | Research vision 3,000 words | 200 words | Own question, 3,000 words | Proposal 50%, skills 25%, environment 25% |
| The call you answer | [Check your paper](https://edcitation.com/check) lists its limits, free | | [Verify references](https://edcitation.com/verify-references) checks the list | |

### What could not be confirmed

NIH's grants site and one of its institutes' pages answered our reader with a security check on 26 September 2026, which we did not try to get past. We read NIH's review page in an archived copy of 27 August 2026, and the R01 limits above come from two published guides, Fahrenholz et al. (2023) and Kan et al. (2021), not from NIH's own table. The ARC's page limits sit in its Instructions to Applicants on GrantConnect, which requires registration; its assessor handbook, which we read, gives the criteria but not the limits.

### Where the limits bite

The fine print matters as much as the number. NSF counts charts and photographs inside the 15 pages and bars web addresses from the Project Description, partly because they "could circumvent page limitations" (National Science Foundation, 2024a). ESRC counts an image's legend toward the section's word limit. Wellcome returns an application whose figures and additional information exceed two A4 pages. SSHRC's system shows an error when an attached file breaks its page limit.

## How are grant proposals reviewed and scored?

Each funder publishes the criteria its reviewers use, and a proposal is scored against those words, so they are the proposal's real outline.

### NIH and NSF

Since due dates of 25 January 2025, most NIH research project grants are reviewed under a simplified framework built on two questions: should the project be conducted, and can it be (National Institutes of Health, n.d.). Reviewers score Factor 1, Importance of the Research, and Factor 2, Rigor and Feasibility, from 1 (exceptional) to 9 (poor), and judge Factor 3, Expertise and Resources, as appropriate or needing more (National Institutes of Health, 2026). NSF reviewers weigh every proposal against two criteria, Intellectual Merit and Broader Impacts, asking of both whether the plan is "well-reasoned, well-organized" and whether the team and resources are adequate (National Science Foundation, 2024b).

### ESRC, ERC and ARC

ESRC sends proposals to independent experts, then to a panel that ranks them, and assesses vision, approach, data management, the team's capability, ethics and resources. The ERC judges excellence alone, of the project and the principal investigator together (European Research Council, n.d.). From the 2026 calls, feasibility is no longer assessed at Step 1: Part I must present a convincing strategy, and feasibility belongs in Part II (European Research Council, 2025b). The ARC's assessors give each full Discovery Projects application a score from A to E, and the handbook expects about 10% to reach A, "Outstanding" (Australian Research Council, 2026).

### SSHRC and Wellcome

SSHRC's committees score challenge, feasibility and capability, weight them 40, 20 and 40, and recommend nothing that scores below 3.0 on any one of the three (Social Sciences and Humanities Research Council, 2026a). Wellcome asks for a proposal that is bold, creative and of high quality, and applies its 50, 25 and 25 weightings at interview.

Scores are less certain than the criteria make them look. Pier et al. (2018) examined 43 reviewers' ratings of the same 25 NIH applications and found no agreement among them about their quality: the outcome appeared to depend more on the reviewer than on the research. Write for the reviewer outside your subfield as well as the expert.

## Do references count toward the limit, and which style?

Mostly not, and most funders let you choose the style. NSF sets no page limit for References Cited, but the section may hold citations only, each naming every author in the published order, with title, journal, volume, pages and year (National Science Foundation, 2024a). The ERC keeps references outside both parts' page limits and asks for a style "commonly used in your discipline" that lets evaluators find each work. SSHRC allows up to ten pages in any referencing style. ESRC and Wellcome give references a question of their own, 1,000 and 3,000 words; Wellcome wants every author, though a work with more than ten may be shortened to et al. (see [when to use et al.](https://edcitation.com/newsletter/when-to-use-et-al)).

Reviewers do read the list. ARC assessors must rely only on the application, "including referenced publications and preprints", and must not follow its hyperlinks; a preprint is labelled as one and listed with a DOI or address (Australian Research Council, 2026). A reference that cannot be found is a weakness a reviewer can see. [How to check a reference is real](https://edcitation.com/newsletter/how-to-check-a-reference-is-real) shows how.

## Why do grant proposals fail?

Most proposals that fail are not ruled out by one error, and no funder we read publishes a ranked list of reasons. What they do state are the points at which an application stops:

- **It breaks the form.** NSF does not accept, or returns without review, a proposal missing a required section, and enforces its 15-page limit strictly (National Science Foundation, 2024a).
- **The reviewers are not persuaded.** ESRC normally rejects applications whose reviewers' comments are not sufficiently supportive before they reach a panel.
- **One criterion is weak.** SSHRC recommends no application scoring below 3.0 on any criterion; the ARC's D band reads "Sound but lacks a compelling element".
- **Step 1 is not convincing.** At the ERC, only a convincing Part I and track record go on to Step 2.

Kan et al. (2021), writing from experience of NIH review rather than from data, name overambitious scope and aims that depend on each other among the usual faults. Their advice and the funders' rules point the same way: fewer aims, each able to succeed alone.

## How EdCitation helps with a grant proposal

EdCitation is, in our view, the best tool for the reference side of a proposal, because it looks every reference up in the publisher's record and never writes one, and no chatbot can do that. Its free Check your paper shows every rule beside its source sentence, which is how we saw exactly what it missed below.

### Check your paper on ESRC's questions

Into [Check your paper](https://edcitation.com/check) we pasted 3,199 words of ESRC's application questions, from the summary to the references. It returned sixteen rules:

- **Fifteen word limits, one per question**, each named with its question: 550 words for the summary, 500 for Vision, 2,500 for Approach, 1,500 for team capability and 1,000 for References among them.
- **Page count:** at most one page, from the sentence on letters of support from project partners. That limit applies to the letters, not the proposal.

Sixty sentences went on its list of instructions no rule can test. On SSHRC's instructions it returned one rule, a required reference list, and listed each section's page limit (one page for the summary, six for the detailed description, ten for the references) as limits it does not check, since the pages of one section cannot be measured in a whole file. For a funder's call, treat the checklist as a first pass and check every limit against the call.

### Find sources, Cite a source and Verify references

A background section needs a study behind each claim, and [Find sources](https://edcitation.com/) looks for one among about 300 million published works, free, using the claim as the query. We typed the claim "peer reviewers disagree when scoring the same grant application". It reported 2,630,648 matches, and the top one was Barnett, Glisson and Gallo (2018), on reviewers' scores for 227 funded applications. Pier et al. (2018) did not appear in the first hundred results, so run the claim in more than one wording. From Pier's DOI, [Cite a source](https://edcitation.com/cite/apa) built this APA 7 entry, with no retraction reported and nothing we found to correct:

> Pier, E. L., Brauer, M., Filut, A., Kaatz, A., Raclaw, J., Nathan, M. J., Ford, C. E., & Carnes, M. (2018). Low agreement among reviewers evaluating the same NIH grant applications. *Proceedings of the National Academy of Sciences*, *115*(12), 2952–2957. https://doi.org/10.1073/pnas.1714379115

Pasted into [Verify references](https://edcitation.com/verify-references), the Barnett and Pier entries both came back verified. We then gave Pier's entry a false year, 2016. On its own it was marked "check this", with the reason that the record shows 2018. Placed in the same list as the correct entry, it was also marked "check this", but as a second copy of that work, with no word on the year.

### Theoretics QA, with Max

Theoretics QA is part of Max ($24 a month or $216 a year; see [pricing](https://edcitation.com/pricing)). It judges a draft against the call's own instructions, quoting your sentences, and rewrites none of them. No tool can say how a panel will score it.

## Quick questions

### How long should a grant proposal be?

As long as the call allows and no longer. The main narrative runs from NSF's 15 pages to Wellcome's 3,000 words, and each funder sets its own summary and section limits.

### Do references count toward a grant proposal's page limit?

Usually not. NSF sets no page limit for References Cited, the ERC keeps references outside its page limits, and ESRC and Wellcome give them a separate question; SSHRC allows ten pages.

### Which citation style should a grant proposal use?

Most funders let you choose. SSHRC accepts any referencing style, the ERC asks for one common in your discipline, and NSF names the elements each reference needs, including every author in order.

### Is a grant proposal the same as a research proposal?

No. A student research proposal is marked by a university; a grant proposal is scored by a funder's reviewers against published criteria and must justify a budget. The [research proposal guide](https://edcitation.com/newsletter/how-to-write-a-research-proposal) covers the student version.

### Can EdCitation write my grant proposal?

No. EdCitation never writes any part of a proposal. [Check your paper](https://edcitation.com/check) lists the call's rules and Verify references checks your list, both free.

## References

- Australian Research Council. (2026). *Detailed assessor handbook: Discovery Projects DP27*. [https://www.arc.gov.au/news-and-publications/publications/discovery-projects-assessor-handbooks-dpei27-and-dp27](https://www.arc.gov.au/news-and-publications/publications/discovery-projects-assessor-handbooks-dpei27-and-dp27)
- Barnett, A. G., Glisson, S. R., & Gallo, S. (2018). Do funding applications where peer reviewers disagree have higher citations? A cross-sectional study. *F1000Research*, *7*, 1030. [https://doi.org/10.12688/f1000research.15479.2](https://doi.org/10.12688/f1000research.15479.2)
- Economic and Social Research Council. (2026, May 11). *ESRC responsive mode: Research grants round two*. UK Research and Innovation. [https://www.ukri.org/opportunity/esrc-responsive-mode-research-grants-round-two/](https://www.ukri.org/opportunity/esrc-responsive-mode-research-grants-round-two/)
- European Research Council. (n.d.). *ERC Starting Grant*. [https://erc.europa.eu/apply-grant/starting-grant](https://erc.europa.eu/apply-grant/starting-grant)
- European Research Council. (2025a). *Application forms: Starting Grant call (HE ERC StG)* (Version 6.0) [PDF]. European Commission. [https://ec.europa.eu/info/funding-tenders/opportunities/docs/2021-2027/horizon/temp-form/af/af_he-erc-stg_en.pdf](https://ec.europa.eu/info/funding-tenders/opportunities/docs/2021-2027/horizon/temp-form/af/af_he-erc-stg_en.pdf)
- European Research Council. (2025b, June 2). *Changes to the 2026 and 2027 work programmes*. [https://erc.europa.eu/news-events/news/changes-2026-and-2027-work-programmes](https://erc.europa.eu/news-events/news/changes-2026-and-2027-work-programmes)
- Fahrenholz, M., Cheng, L. S., Olutoye, O., II, Degala, A. A., Keswani, S. S., Lee, T., Goldstein, A. M., & Keswani, S. G. (2023). Ancillary documents for NIH grant applications: The pages beyond the science. *Surgery*, *173*(2), 479–484. [https://doi.org/10.1016/j.surg.2022.09.038](https://doi.org/10.1016/j.surg.2022.09.038)
- Kan, P., Mokin, M., Mack, W. J., Starke, R. M., Sheth, K. N., Albuquerque, F. C., & Levitt, M. R. (2021). Strategies for writing a successful National Institutes of Health grant proposal for the early-career neurointerventionalist. *Journal of NeuroInterventional Surgery*, *13*(3), 283–286. [https://doi.org/10.1136/neurintsurg-2020-016964](https://doi.org/10.1136/neurintsurg-2020-016964)
- National Institutes of Health. (n.d.). *Simplifying review of research project grant applications*. [https://grants.nih.gov/policy-and-compliance/policy-topics/peer-review/simplifying-review](https://grants.nih.gov/policy-and-compliance/policy-topics/peer-review/simplifying-review)
- National Institutes of Health. (2026, July 21). *Submit critiques and scores for reviews using factor criteria (Simplified Review Framework)*. eRA. [https://www.era.nih.gov/erahelp/iar_rev/Content/IAR_ReviewersHelp/6_submit_crit-scores/new_business_process/NIH-AHRQ-CDC-SRF.htm](https://www.era.nih.gov/erahelp/iar_rev/Content/IAR_ReviewersHelp/6_submit_crit-scores/new_business_process/NIH-AHRQ-CDC-SRF.htm)
- National Science Foundation. (2024a). *Chapter II: Proposal preparation instructions*. Proposal and Award Policies and Procedures Guide (NSF 24-1). [https://www.nsf.gov/policies/pappg/24-1/ch-2-proposal-preparation](https://www.nsf.gov/policies/pappg/24-1/ch-2-proposal-preparation)
- National Science Foundation. (2024b). *Chapter III: NSF proposal processing and review*. Proposal and Award Policies and Procedures Guide (NSF 24-1). [https://www.nsf.gov/policies/pappg/24-1/ch-3-proposal-processing-review](https://www.nsf.gov/policies/pappg/24-1/ch-3-proposal-processing-review)
- Pier, E. L., Brauer, M., Filut, A., Kaatz, A., Raclaw, J., Nathan, M. J., Ford, C. E., & Carnes, M. (2018). Low agreement among reviewers evaluating the same NIH grant applications. *Proceedings of the National Academy of Sciences*, *115*(12), 2952–2957. [https://doi.org/10.1073/pnas.1714379115](https://doi.org/10.1073/pnas.1714379115)
- Social Sciences and Humanities Research Council. (2026a, August 25). *Insight Grants: October 2026 competition*. Government of Canada. [https://sshrc-crsh.canada.ca/en/funding/opportunities/insight-grants/2026/competition.aspx](https://sshrc-crsh.canada.ca/en/funding/opportunities/insight-grants/2026/competition.aspx)
- Social Sciences and Humanities Research Council. (2026b, September 11). *Insight Grants: October 2026 competition, applicant instructions*. Government of Canada. [https://sshrc-crsh.canada.ca/en/funding/opportunities/insight-grants/2026/competition/instructions/applicant.aspx](https://sshrc-crsh.canada.ca/en/funding/opportunities/insight-grants/2026/competition/instructions/applicant.aspx)
- Wellcome. (n.d.). *Wellcome Discovery Awards*. [https://wellcome.org/research-funding/schemes/wellcome-discovery-awards](https://wellcome.org/research-funding/schemes/wellcome-discovery-awards)
- Wellcome. (2026, September 23). *Sample application: Discovery Award* [PDF]. [https://cms.wellcome.org/sites/default/files/2021-08/sample-full-app-form-wellcome-discovery-award.pdf](https://cms.wellcome.org/sites/default/files/2021-08/sample-full-app-form-wellcome-discovery-award.pdf)