NSF Letter of Collaboration: What It Must (and Must Not) Say

An NSF letter of collaboration is a short, formally restricted document that documents an unfunded, uncompensated collaborative arrangement on an NSF proposal — someone contributing meaningfully to the proposed research without being included in the project budget. Because NSF’s Proposal & Award Policies & Procedures Guide (PAPPG) tightly limits what these letters can say, getting the content and format wrong is a common, avoidable compliance flag: NSF explicitly prohibits language that a research administrator might otherwise consider standard for this kind of letter, including any commitment of organizational resources or any endorsement of the project’s merit.

What an NSF Letter of Collaboration Documents

NSF’s PAPPG addresses letters of collaboration under Chapter II’s Special Information and Supplementary Documentation instructions, as part of “Documentation of Collaborative Arrangements of Significance to the Proposal.” The letter exists for one narrow purpose: to confirm that a named individual outside the proposal’s budget intends to participate in the project if it is funded. It is not evidence of merit, and it is not a funding or resource commitment — it is a factual confirmation that a real collaborative relationship exists.

This matters because NSF, unlike NIH, does not treat letters attached to a proposal as advocacy documents. NSF reviewers are instructed to disregard any letter content that reads as an endorsement, and program officers can flag a proposal where a collaboration letter oversteps its narrow evidentiary role.

When NSF Expects a Letter of Collaboration

A letter of collaboration is appropriate when a proposal describes substantial collaboration with an individual — often at another institution, agency, or organization — who is contributing meaningfully to the intellectual content of the project but who is not a paid member of the project’s key personnel and does not appear in the budget. Common examples include:

• An external researcher who will provide specialized expertise, data, or analysis but is not requesting NSF funding through this proposal.
• A collaborator at another institution who will contribute effort in kind, without compensation from the award.
• An individual whose participation the Project Description or the Facilities, Equipment and Other Resources section already describes, and who needs to confirm in writing that the description of their involvement is accurate.

If the collaboration instead involves a subaward, consultant agreement, or any budgeted cost, the correct documentation is a budget justification and, where applicable, a subaward budget or consultant letter — not a letter of collaboration. NSF reviewers can tell the difference between an unfunded collaborator’s letter and an attempt to route what is functionally a subaward commitment through an unbudgeted letter, and conflating the two is itself a compliance problem.

Letter of Collaboration vs. Letter of Support vs. Letter of Intent

These terms are frequently used interchangeably by proposal writers, but at NSF they are not interchangeable, and using the wrong one can create exactly the kind of overreaching language the PAPPG prohibits:

• Letter of collaboration (the NSF term of art): confirms an unfunded collaborator’s intent to participate if the proposal is funded. Content is deliberately minimal — intent to participate, nothing more.
• Letter of support: NIH’s umbrella term for a broader category of letters confirming a resource, role, or commitment from a contributor. NSF proposals sometimes use this label loosely for the same document a letter of collaboration is meant to be, but NSF’s own guidance consistently uses “letter of collaboration” for this specific unfunded-participation case and restricts its content more tightly than NIH restricts an equivalent letter. See CASRAI’s guide to letters of support for a grant proposal for how the two major US funders diverge on this document category more broadly.
• Letter of intent: an entirely different document at a different stage of the process — a short pre-proposal notice, submitted before a full application, telling a funder or program that an investigator plans to submit a full proposal. NIH uses letters of intent this way for many funding opportunity announcements; NSF’s use is program-specific rather than a general requirement. A letter of intent has nothing to do with confirming a collaborator’s role and should not be confused with a letter of collaboration.

A related but distinct document is the letter of commitment, used when a solicitation requires proof that a specific person or resource — often a mentor, a facility with limited capacity, or key personnel — will actually be available at the level claimed in the proposal. A letter of commitment is about certainty of availability; a letter of collaboration is about confirming an unfunded working relationship. NSF program solicitations occasionally require one, the other, or both, so the specific solicitation’s instructions govern which document (or documents) a given collaborator should provide.

What the PAPPG Prohibits: Content and Format Rules

The PAPPG’s restriction on letters of collaboration is one of the more strictly enforced formatting rules in NSF proposal preparation, precisely because it is easy to get wrong by analogy to NIH practice. Three rules govern the content:

• No endorsement of intellectual merit. The letter must not evaluate, praise, or advocate for the proposed project — that is the reviewers’ job, and NSF treats a letter that does this as improperly attempting to influence merit review.
• No commitment of resources or cost-sharing. A letter of collaboration cannot pledge funding, equipment, facility access, or any other organizational resource on behalf of the collaborator’s institution. If a real resource commitment is being made, that belongs in the Facilities, Equipment and Other Resources section of the proposal itself (and, where cost-sharing is involved, is subject to NSF’s separate and more restrictive cost-sharing policy) — not in a collaboration letter, which NSF wants limited to a statement of intent to participate.
• No duplication of budgeted information. If the collaborator’s role and compensation are already documented through the budget (for example, as a subaward or consultant), that information follows the budget-documentation procedures elsewhere in the PAPPG rather than being restated in a collaboration letter.

NSF’s own recommended language for the letter is intentionally narrow. A commonly used, NSF-consistent formulation (adapted by many university sponsored-programs offices from PAPPG guidance) reads approximately:

“If the proposal submitted by Dr. [Principal Investigator’s full name] entitled [proposal title] is selected for funding by NSF, it is my intent to collaborate and/or commit resources as detailed in the Project Description or the Facilities, Equipment and Other Resources section of the proposal.”

The PAPPG’s own framing of the restriction is direct: letters of collaboration “should be limited to stating the intent to collaborate and should not contain endorsements or evaluation of the proposed project.” Sponsored-programs offices routinely reject draft letters that add laudatory language (“this is an outstanding and highly innovative proposal…”) or that promise specific resources, and ask the collaborator to resubmit a compliant version before the proposal is finalized.

Format checklist

• One letter per unfunded collaborator, on that collaborator’s institutional letterhead, dated and signed by the collaborator (not by the PI on the collaborator’s behalf).
• Addressed generically to NSF or the relevant program, not written as marketing copy.
• Uploaded as Supplementary Documentation in the proposal submission system (Research.gov or Grants.gov, per the specific proposal type), not embedded in the Project Description.
• Unique to the specific proposal being submitted — a letter recycled across multiple submissions without update, or altered by anyone other than the original signatory without the signatory’s approval, does not meet NSF’s expectations for this document.
• Short. NSF does not specify a hard page limit for collaboration letters, but because the permitted content is a single statement of intent, a compliant letter is typically well under one page.

Common Mistakes That Get a Letter of Collaboration Flagged

• Writing it like an NIH letter of support. Language that would be normal and welcome in an NIH letter — “Dr. [Name]’s proposal represents an important advance in the field, and I am pleased to support it” — reads as exactly the kind of merit endorsement NSF’s PAPPG prohibits.
• Promising resources instead of participation. “Our lab will provide access to the XYZ instrument at no cost” belongs in Facilities, Equipment and Other Resources, described by the proposing institution, not asserted as a commitment inside a collaboration letter.
• Using a collaboration letter to route what is really a subaward. If the collaborator’s institution is receiving any portion of the NSF award, that relationship needs a subaward budget and justification, not a letter of collaboration.
• Reusing the same letter across proposals without updating it. The PAPPG expects each letter to be specific to the proposal it accompanies.
• Missing signature or letterhead. A collaboration letter is a real document from a real collaborator — an unsigned draft, or one clearly drafted entirely by the PI and merely forwarded for signature without genuine review, undermines the evidentiary purpose the letter is meant to serve.

Frequently Asked Questions

Is an NSF letter of collaboration required on every proposal?

No. It is only needed when the proposal describes substantial, uncompensated collaboration with someone outside the funded project team. Many NSF proposals have no unfunded collaborators and therefore need no letter of collaboration at all. Always check the specific program solicitation, since some programs add their own requirements on top of the general PAPPG rule.

Can a letter of collaboration mention specific equipment or facilities the collaborator will provide?

It can describe the collaborator’s role, but it should not be the vehicle for committing resources — that belongs in the proposal’s Facilities, Equipment and Other Resources section, written and controlled by the proposing organization, not asserted as a pledge inside the collaborator’s letter.

What happens if a letter of collaboration includes prohibited language?

NSF reviewers are instructed to disregard improper endorsement language, but a program officer or the reviewing panel may still flag the proposal, and many institutional sponsored-programs offices will catch and reject a non-compliant draft before submission as part of internal proposal review.

Does the PI write the letter of collaboration?

The PI or the proposing institution’s sponsored-programs office often drafts a compliant template for the collaborator to review, but the letter must be signed by the collaborator on their own institutional letterhead — it needs to represent the collaborator’s own confirmation, not be presented as if unilaterally written by the applicant.

How is a letter of collaboration different from a letter of commitment?

A letter of commitment confirms that a specific person or resource will actually be available at the level of effort or access the proposal claims — common for training-grant mentors or capacity-limited facilities. A letter of collaboration confirms an unfunded collaborative relationship exists and that the collaborator intends to participate if funded. Some NSF solicitations require one, the other, or both; the solicitation’s own instructions govern.

References

• National Science Foundation, Proposal & Award Policies & Procedures Guide (PAPPG), NSF 24-1, Chapter II: Proposal Preparation Instructions — Special Information and Supplementary Documentation, Documentation of Collaborative Arrangements of Significance to the Proposal through Letters of Collaboration.
• National Science Foundation, PAPPG (NSF 24-1) overview — effective for proposals submitted or due on or after May 20, 2024; supplemented since by targeted Policy Notices (Supplement 1, NSF 26-200, and Supplement 2, NSF 26-202) rather than a renumbered edition.