NDSEG Research Proposal: Format and Redaction Guide

NDSEG Research Proposal: Format and Redaction Guide

The NDSEG research proposal is one PDF of no more than four pages, and you upload it twice. The application site gives the format in its own words: "4 Page Proposal with work cited", made of a "3 Page Proposal Single line space" plus "1 Page Max Cited Work", set in "12Pt. Times New Roman Font-a smaller font size may be used for cited work, charts and graphics", with "Margins should be no less than 0.5". "Graphics and Charts may be included", and "All Files must be uploaded as PDFs ONLY". The upload is listed as "Research Proposal [Redacted and Original Versions]".

The redaction rule is where applicants do real damage, and the application FAQ defines it precisely:

"Redaction is defined as no inclusion of any identifying information, such as your or any person(s) names, organization/institution names, or specific details that could tie the application back to you and any institution/organization or person. Failure to follow this guidance will prevent your application from moving forward in the evaluation process."

Every page that ranks for this fellowship says the proposal must be redacted. Almost none says how — so applicants delete the science along with the names. Below: what to remove, how the two files are named, and the move that keeps your methods and results intact once the proper nouns are gone. The two typed essay boxes run on separate rules; see the NDSEG fellowship essays guide.

Table of Contents

• The format contract
• Two files, one proposal
• What the redacted proposal must not contain
• The black-bar failure, and the move that fixes it
• Before you upload: verifying the file
• DoD alignment without compliance theater
• The citations page
• Format numbers that circulate and are wrong
• Advisor first, review alongside
• Frequently Asked Questions

The format contract

The capitalization and spacing above are the sponsor's, quoted as published.

The applicant tips settle the page-count question: "Keep in mind the actual proposal is no more than 3 pages, the 4th page is for the cited work only". That is one PDF, four pages at most — not a three-page file plus a separate citations file.

One published fact names your reader: applications are "evaluated by a panel of academic experts who have doctoral degrees from the science and engineering discipline of the applicant's proposed advanced degree program." A discipline-matched expert, reading a document with your institution removed.

Two files, one proposal

The component list shows the same pattern twice: "Research Proposal [Redacted and Original Versions]" and "Resume/CV [Redacted and Original Versions]". Two documents, four files.

The naming convention comes from the application's own upload instructions, as recorded in the OMB-approved application instrument. Those screens were captured from an earlier cycle's portal, so treat the wording as the instrument record rather than text read off the current site; the live pages independently state both the dual upload and the redaction definition. From the record:

"Research Proposal [Unredacted Original Version]: Please upload your research proposal in PDF Format using the following file name format: Lastname_Firstname_ResearchProposal"

"Uploaded redacted research proposal: Please upload a redacted research proposal that does not include University Name, Organization/Laboratory Name, Fellow or Other Names, No Logos or Photos of individuals or Fellow Upload your research proposal in PDF Format using the following file name format: Redacted_ResearchProposal"

The file name is part of the redaction: exporting the redacted version under your own name puts your surname back on the document you just scrubbed.

The same screen carries a separate rule, about completeness rather than redaction: "If for any reason an application does not contain all of the required elements, it will NOT be reviewed. It is the applicant's responsibility to ensure the completeness of their application."

Redaction also works differently in different places — the proposal PDF, the CV PDF, and the essays typed into the form. The essay side is covered in the essays guide and reviewable against the NDSEG application essays rubric.

What the redacted proposal must not contain

From the upload instructions, the removal list is five items: university name; organization or laboratory name; fellow or other names; logos; photos of individuals.

That is the easy half. The harder half is the FAQ's catch-all — "specific details that could tie the application back to you and any institution/organization or person."

State the boundary honestly, because you will be making judgment calls under it. A careful read catches visible identifiers: names, explicit affiliations, a lab's public project title, a sentence naming the facility. Nothing and nobody can certify that an ordinary technical detail fails to identify you to a specialist in your subfield — real-world anonymity is not something a document check confirms. Treat the removal list as the floor, and "would three people in my area know whose proposal this is?" as an advisor question.

Note what the rule does not say. It does not ban technical terminology, instrument classes, publicly known method names, award categories described generically, or the title of the DoD strategy document you aligned to.

Two structural facts make the boundary less frightening than it looks. First, the form separately collects your research advisor's name, institution, department, email, and phone number. Redaction is a document-level rule for the uploaded PDFs, not an application-level secrecy regime — taking your advisor's name out of the proposal costs the program nothing it does not already have. Second, the form asks you to agree that the proposal "will be reviewed by Subject Matter Experts in your discipline including those from Academia, Industry and Government," and asks separately for a "Publicly Releasable Abstract". Results your group is not ready to share do not belong in a document designed to circulate.

The black-bar failure, and the move that fixes it

The failure looks like this. You delete the university, the lab, the advisor, the collaborators, the funded project name. So far, correct. Then momentum takes over: out goes the beamline, the instrument, the cohort, the sample size, the decision rule, the number. What reaches the panel is a topic statement with no owner and no evidence — the redaction removed the identity and the science, and only one of those was required.

The move that prevents it is role-bounded generalization. Keep every action, method, decision, parameter, and outcome. Replace each identifier with the smallest true role-or-setting class that still lets a discipline-matched expert judge the work.

The sponsor models this itself. Asked how to handle an award you cannot name, the FAQ answers that you "can give the general description of the award without mentioning the organization name. For example, you can say 'I received a prestigious university award given to students with outstanding academic achievement…'" The organization disappears; the meaning survives.

Constructed teaching material — invented for illustration, with no real institution, laboratory, advisor, or grant.

The same idea at paragraph length. Constructed teaching material, invented for this guide — not an excerpt from any real proposal, and no real winning proposal should be copied from any source.

Original, identity-laden: "Building on the solid-state electrolyte program in Professor [Name]'s group at [University], using beamline time from our collaboration with [Laboratory], I will extend the composition series our group published last year."

Naive redaction, gutted: "Building on prior work, and using specialized equipment, I will extend earlier research on battery materials. This work is important for energy storage."

Role-bounded redaction: "A university group recently reported dendrite penetration in sulfide solid-state electrolytes above a critical current density, without resolving whether the threshold is set by bulk conductivity or grain-boundary chemistry. I will synthesize four compositions along a substitution series, measure critical current density by galvanostatic cycling at three temperatures, and image failed cells at a synchrotron beamline with sub-micron resolution to locate initiation sites. If segregation sets the threshold, critical current density will track boundary composition rather than bulk conductivity."

The third version names no institution, advisor, facility, or project. It does have a question, a hypothesis, a method, a measured quantity, and a result that would count as evidence either way.

The test for every redacted sentence: does a discipline-matched expert still know what will be done, to what, how, and what result would count? If the only thing the sentence lost was a proper noun, that is exactly right.

Before you upload: verifying the file

The FAQ names the tools — "Recommend you utilize MS Word black highlighter feature or Adobe redaction feature." Then verify the exported file rather than trusting the editing step:

• Try to select and copy the text under each black bar. A drawn rectangle hides pixels; the characters may still be selectable.
• Search the PDF for your surname, your advisor's surname, your institution, your laboratory, your department, and your city.
• Check the document properties, which can carry an author or company name from your word processor.
• Inspect every figure at full zoom for a surviving logo, watermark, or photograph of a person.
• Check the file name: Redacted_ResearchProposal.

That is a verification procedure you run, not a security guarantee about any tool. Leave time for it — the cycle closes on October 30, 2026 at 5 PM Eastern, and a second version of a technical document is a multi-session job.

DoD alignment without compliance theater

The most expensive misreading of this application is treating the linked Broad Agency Announcements as a specification to satisfy. The FAQ is unambiguous:

"You are NOT required to fulfill any of the requirements put forth by the DoD Strategy Documents. You are only using them as reference to align your research and proposal to DoD strategic and scientific research mission, initiatives, and priorities."

The DoD agencies page says the same from the other side: the BAAs are "informational only", and "You will use the BAAs only to determine research topics relevant to DoD interests."

The application's structure explains why. The DoD strategy document title, your project title, your single-select agency preference, and a publicly releasable abstract are all form fields filled in outside the proposal. The alignment claim is already captured. So the body's job is not recited solicitation language, section numbers, or compliance phrasing — it is one visible scientific connection: what your objective, method, or expected output has to do with the research area you named. One sentence a panelist can point to beats a paragraph of borrowed vocabulary.

Two cautions. The agency preference is single-select, and the form states that "Selections will not be based on your preference, but the needs of the DoD agency" — naming a branch is not a lever. And the no-contact rule is absolute: "PLEASE DO NOT CONTACT THE PROGRAM MANAGERS listed in the BAAs for any reason." The same guidance adds, "If you require assistance, please work with your advisor(s)." Questions go to your advisors, and administratively to NDSEG.PMO@sysplus.com.

The 2023 National Defense Science and Technology Strategy named fourteen critical technology areas, from microelectronics and quantum science to biotechnology and hypersonics — a dated 2023 list, not a closed or current eligibility menu. Open the documents currently linked from the DoD agencies page instead, and keep the claim bounded: a proposal promises a contribution to knowledge in a named area, not deployment, fielding, or a national outcome.

The citations page

Page four, one page maximum, cited work only, with smaller type explicitly permitted there and in charts and graphics. The application publishes no required citation style, so the burden is consistency: pick one and hold it.

What nothing else on this topic addresses is that your reference list is a redaction surface. A bibliography carrying your own first-author papers, your advisor's name, or your group's preprints re-identifies you as efficiently as letterhead. In order:

1. Prefer citations that do not identify you. Most claims can be grounded in the wider literature rather than your group's output.
2. Where you must cite identifying work, describe the finding and attribute it generically. The sponsor's own model: "You can mention the title but do not mention authors/co-authors and name of the journal or organization it was published under."
3. If you are unsure whether an entry identifies you, ask your advisor — not the program office, and not a program manager.

One more failure, common in proposals written at speed: references sit on page four, but no sentence in the body ties a specific claim to a specific source. Each entry should be load-bearing for something a reader could challenge.

Format numbers that circulate and are wrong

A retired official NDSEG domain still ranks well in search and still publishes the previous cycle's rules. Those numbers are superseded, not an alternative view:

If a page shows an October 2022 deadline, 11-point type, 1-inch margins, or a separately uploaded citations file, it is describing an older cycle. Check every number against the program's own application site — and check third-party templates too, since a template can encode retired rules invisibly.

Advisor first, review alongside

The program tells you to do this: "Have an Advisor review your Proposal, Personal Statement/Application", and "If you require assistance, please work with your advisor(s)." Your advisor owns the judgments nobody else can make — whether the project is feasible, whether the scientific claim holds, and whether a retained detail identifies you inside your subfield.

A rubric review is the repeatable pass you run between drafts, alongside that advisor read rather than in place of it. On the redacted version you paste in, the NDSEG research proposal rubric reads for whether each objective has a matching method and a stated evidence type, whether the DoD connection is stated in a sentence rather than implied by an acronym, whether the redaction left actions and outcomes intact, and whether identifiers are still visible in what you supplied.

The boundary, plainly: a review reads only the version you paste. It cannot see your original, compare the two versions, inspect margins, type size, page count, PDF export, logos, or file names, or certify that a document is anonymous.

More STEM fellowship guides

• NDSEG fellowship essays guide — the two typed 500-word boxes and the cycle window
• STEM PhD fellowship portfolio: GRFP, NDSEG, Hertz — the deadline ladder and what transfers
• NSF GRFP cycle guide — the contrasting genre, with its own limits, required headings, and no redaction rule; reviewable against the GRFP research plan rubric
• How to write a scholarship essay that fits the award
• How a scholarship review works

Review your draft

When the redacted version is written, paste it into the scholarship essay review and run it against the NDSEG research proposal rubric. Two Quick Reviews are free each day; your first Full Review is $5, once per account, and Pro is $50 for ten Full Reviews. Reviews are of your own writing only, and feedback is a revision signal rather than a verdict on your application.

Frequently Asked Questions

Do I really upload the proposal twice?

Yes. The application lists "Research Proposal [Redacted and Original Versions]", and the upload instructions name each file: the original as Lastname_Firstname_ResearchProposal, the redacted version as Redacted_ResearchProposal. The CV follows the same pattern.

Can I keep my advisor's name if the form already collects it?

No — take it out of the proposal PDF. The form separately collects your advisor's name, institution, department, email, and phone number, so removing it from the document costs the program no information.

Do I have to meet the requirements in the BAA I aligned to?

No: "You are NOT required to fulfill any of the requirements put forth by the DoD Strategy Documents." And do not contact the people named in them — "PLEASE DO NOT CONTACT THE PROGRAM MANAGERS listed in the BAAs for any reason." Work with your advisors, and use NDSEG.PMO@sysplus.com for application questions.

How long is the proposal, and does the citations page count?

One PDF of no more than four pages: three pages of body, single line spacing, plus one page maximum of cited work. The citations page is the fourth page of that document, not a separate upload.

What happens if I get the redaction wrong?

Failure to follow the redaction guidance "will prevent your application from moving forward in the evaluation process." A separate rule covers completeness: if an application "does not contain all of the required elements, it will NOT be reviewed."

Sources

• NDSEG — Application, Evaluation & Award — format contract, dual upload, applicant tips, evaluation panel
• NDSEG — Application FAQs — redaction definition and consequence, redaction tools, BAA requirements, contact policy
• NDSEG — DoD Supporting Agencies — BAAs as informational only, no-contact rule
• NDSEG — How to Apply and program home — FY2027 window
• OMB information collection record (instrument PDF) — upload instructions, file naming, completeness clause, form fields
• 2023 National Defense Science and Technology Strategy — critical technology areas. The strategy document is linked from the sponsor's own NDSEG DoD Supporting Agencies page, which is the navigation path to use if the direct file link does not open