Grant applications and the rules that come before the science

What are the formatting rules, and do they really matter?

They are printed, and the consequence of missing them is printed alongside. The National Institutes of Health asks for text at "11 points or larger" and margins of "at least one-half inch" on all four sides, with paper size no larger than 8.5 by 11 inches.¹ It requires PDF, and asks that attachments carry "descriptive filenames of 50 characters or less (including spaces)."¹

The consequence sits in the same paragraph. Applications with attachments that do not conform to those minimums "may be withdrawn from consideration."¹

The National Science Foundation sets different numbers for the same items. It accepts Arial, Courier New, or Palatino Linotype at 10 points or larger, Times New Roman at 11 points or larger, and the Computer Modern family at 11 points or larger.² Its density rule is "no more than six lines of text within a vertical space of one inch," and margins "in all directions, must be at least an inch."²

The two agencies disagree on nearly every number, which is why a proposal reformatted from one to the other has to be checked item by item.

Where do the page limits come from?

From the funding opportunity first and the general guide second. The NIH page limits table applies "unless otherwise specified in the notice of funding opportunity."³ Its own statement of precedence is explicit: "Funding opportunity instructions always supersede general application guide instructions and NIH Guide notice information supersedes both the funding opportunity and the application guide."³

Limits vary by the kind of award rather than by the field. The Research Strategy runs to five pages for a DP1, six for an R03 or R21, ten for a DP2, and twelve for an R01 and most other activity codes.³

NSF sets a single figure and defends it. The Project Description "may not exceed 15 pages," and that count includes "visual materials, including charts, graphs, maps, photographs, and other pictorial presentations."² Conformance "will be strictly enforced," and NSF states it will not accept or will return without review proposals inconsistent with the instructions.²

What kind of edit does a grant application need?

A copyedit and a consistency pass, run after the science is settled. Editors Canada's definition of copy editing covers correcting grammar, spelling, punctuation, and usage while checking for consistency and continuity.⁴ That is the level most applications need.

Applications also carry a structural problem editing can reach. The same study is described in the abstract, the specific aims, the research strategy, and the budget justification, usually by different hands and on different days. Aims that are numbered three in one document and lettered in another cost a reviewer time.

What an editor does not do is decide whether an aim is feasible, whether a method suits a question, or whether a preliminary result supports a hypothesis. Those belong to the investigators and to the people who review the application.

What does an editor need from your side before starting?

The notice of funding opportunity itself, since it overrides the general guide. The list of attachments and the limit on each. The name of the person who has authority over the wording. Finally, the deadline as your sponsored programs office understands it, which is often earlier than the agency's.

Anything already cleared elsewhere is worth marking. Biographical sketches, facilities descriptions, and institutional boilerplate frequently come from a research office and are not yours to change.

Where figures and their captions are still moving, saying so is worth a sentence in the brief. It prevents an editor from making a caption consistent with a figure that is about to be replaced.

The style convention matters here more than in most business writing. Reference format, abbreviation of journal titles, and the treatment of gene and organism names follow discipline conventions rather than a general manual, and naming the convention avoids a round of unnecessary changes.

Can an application be cut to the page limit without losing the science?

Sometimes, and the honest answer depends on how far over it is. Cutting a page from thirteen is usually a matter of removing repetition between sections, tightening sentences that restate their own subject, and moving detail into a figure legend where the rules allow it.

Cutting four pages is a different request. That is a decision about which content leaves, and the investigators make it. An editor can mark every passage that appears twice and every sentence that carries no information a reviewer needs, then hand the list back.

Formatting is not the place to find the space. Both agencies set type size and margins as minimums precisely because applicants have tried, and NSF notes that its limit may not be exceeded unless a deviation has been specifically authorized.²

When in the process should the edit happen?

After the science is stable and before the internal deadline, with several days between the two. An edit delivered the morning of submission cannot be acted on, since accepting changes across a dozen attachments takes longer than people expect.

Institutional review usually adds a stage. Many research offices require the application some days ahead, and an edit that finishes after that point has to be squeezed into whatever time the office allows for corrections.

Where a resubmission is planned, the introduction responding to the previous review is the piece most worth an editor's attention. It is short, it is read closely, and it is usually written last, which is a poor combination.

Applications assembled from an earlier submission carry the previous version's inconsistencies forward. A comparison pass against the old file catches the aim that was renumbered and the collaborator who left.

Who decides on a disputed change?

The principal investigator. Every change arrives as a tracked change with a comment where a reason is needed, and any of them can be rejected without discussion.

Disagreements on applications generally involve field vocabulary rather than grammar. Where a phrase reads oddly to a general reader but is standard usage in the discipline, the discipline wins. A marginal query is the way to establish which of the two cases applies, since only the investigators know.

Multi-investigator applications need one arbitrator named in advance. Contributors who each hold a different view of a shared paragraph cannot be reconciled by an editor.

Every document a client sends us is treated in confidence, which includes unpublished preliminary data and anything under embargo. Applications frequently contain results a year away from publication, and they stay with the people who produced them.

Can an editor make an application fund?

No, and the funding rate makes that plain enough without anyone claiming otherwise. Awards turn on significance, approach, investigators, environment, and the judgment of a study section, none of which an editor influences.

What editing removes is the avoidable reason for a low score. A reviewer who has to read a sentence twice, or who cannot find the aim being discussed, is spending attention on the writing instead of the work.

Applications withdrawn for a formatting fault never reach that stage at all, and that is the part of the problem with a fixed and knowable answer. The rules are published, the numbers are specific, and checking a submission against them takes an afternoon rather than a week.

An application that is correct on its face has not been made more likely to succeed. It has been kept in the pool where success is possible.