AgriXiv All articles
Publishing & Policy Analysis

The Sixth Tab: How Federal Grant Guidance Learned to Account for Preprints

AgriXiv
The Sixth Tab: How Federal Grant Guidance Learned to Account for Preprints

Photo by Photo by Vitaly Gariev on Unsplash on Unsplash

The spreadsheet had six tabs. Five existed because the application form demanded them: personnel, budget justification, facilities, current and pending support, and a publications list that stopped at ten entries. The sixth existed because Dr. Priya Raghunathan, an assistant professor of soil fertility at a land-grant university in southern Minnesota, had learned to keep a private record of everything that did not fit anywhere else.

By February 2026, that tab held eleven rows. Nine were conventional journal articles. Two were preprints posted to AgriXiv, both of them revisions of work she had presented at a regional agronomy meeting, neither of them yet accepted anywhere. When she submitted her application to the Agriculture and Food Research Initiative's foundational program, she cited both. Then she spent three weeks worrying about it.

Where the Language Actually Lives

Ask a program officer whether preprints are allowed in a NIFA application and the answer is yes, with qualifications. Ask where that is written down and the conversation becomes a tour of the agency's document architecture.

The agency's programmatic terms and conditions address research outputs in general terms, and its guidance on products asks applicants and awardees to distinguish reviewed from unreviewed work. Individual program solicitations sometimes add a sentence of their own; often they say nothing. The Agricultural Research Service, which operates under different authorities and employs its own scientists rather than making grants, has handled the question separately, with internal review requirements for staff who wish to post working papers before or during journal review. The result is not a prohibition. It is closer to an accounting convention, applied unevenly, in which a preprint is permitted but rarely counted.

"The safe reading is that preprints are permitted but are not peer-reviewed publications," said a national program leader who asked not to be named because she was describing internal practice rather than official policy. "The harder question is what they count as. Right now they count as other."

The Reviewer Who Asked for a DOI

Raghunathan's study section raised no objection to the citations themselves. What it raised was a question about provenance. One reviewer noted that the second preprint had been revised twice on the server and asked whether the application cited the version of record or an earlier posting. It was a fair question with an awkward answer: she had cited version two in the narrative and version three in the bibliography, because the revision had posted after she wrote the first draft and before she assembled the second.

She fixed it in a resubmission. The episode cost her an afternoon and, she believes, nothing else. But it illustrates a mechanical problem that agencies have been slow to address. Preprints are versioned objects, and a grant file is a document that freezes. Every citation to a working paper carries a small obligation to say which one.

"Nobody in that review panel was hostile to the idea," she said. "They just wanted to know what they were reading. That is not a controversy. That is a copyediting problem that the system hasn't solved yet."

What the Final Report Cannot Say

The more instructive friction came after the award, in the section of the annual report devoted to products. The form offers categories: journal articles, books, conference papers, patents, germplasm releases, datasets, other. A preprint has no obvious home. Awardees have developed folk conventions, filing them under conference papers, or under other with a note, or omitting them and mentioning them in the narrative, where they are read by fewer people.

This matters beyond bookkeeping. Products feeds the agency's public accountability reporting, and it feeds the internal metrics by which early-career researchers are sometimes judged when they return to their own institutions. A postdoctoral researcher who posts four working papers in a year, all of them eventually published in decent journals, may look less productive on an agency product table than a colleague with one slower article and a conference abstract.

Raghunathan's department chair, asked whether the tenure committee would count the preprints, gave an answer that has become familiar across the land-grant system: they would count the resulting articles, and they would view the preprints as evidence of activity rather than as products in themselves. Which is to say the server is treated as a workshop and not a press.

The Argument for Clearer Rules

None of this is a scandal. It is what policy looks like when practice moves first. Preprints have become ordinary in agronomy, plant pathology, and agricultural economics, and the agencies' documents have begun to acknowledge them without yet deciding what they are for.

Three changes would cost little. First, a single agency-wide sentence defining a preprint for grant purposes and stating plainly whether it may be cited, rather than leaving applicants to triangulate. Second, a product category for unreviewed working papers, so that the reporting matches the reality. Third, a requirement, or at least a strong encouragement, to cite by version and date, which would resolve the provenance question without anyone having to guess.

There is a fourth, harder one: whether preprints should count as evidence of productivity for the early-career scientists the agency is explicitly trying to support. That is not a paperwork matter. It is a judgment about what open research is worth, and the agencies have not been asked to make it out loud.

Raghunathan keeps the sixth tab anyway. In October she added a twelfth row, a working paper on cover crop termination timing, and cited it in a renewal application without much anxiety. "The rules are fuzzy," she said, "but they are fuzzy in a direction that is more permissive than they were five years ago. I would rather have ambiguity than a ban."

That is a reasonable position for an individual researcher. It is a thin foundation for a funding system that now spends real money on work that first appears, deliberately, before anyone has reviewed it.

All Articles

Related Articles

Thirty-Four Citations, One Without a Journal: How a Rangeland Carbon Preprint Reached BLM Guidance

Thirty-Four Citations, One Without a Journal: How a Rangeland Carbon Preprint Reached BLM Guidance

When an Unreviewed Paper Takes the Stand: Irrigation Science in a Western Water Court

When an Unreviewed Paper Takes the Stand: Irrigation Science in a Western Water Court

The Market Reads First: An Early Yield Paper and the WASDE Calendar

The Market Reads First: An Early Yield Paper and the WASDE Calendar