Skip to content
Scriptorium Scriptorium

NIH significance patterns: what funded grants look like

Last updated: 2026-10-20

Synthesis

Section titled “Synthesis”

NIH grant applications — particularly R01s — have a recognizable structural anatomy that has crystallized over decades of practitioner experience and is reinforced by the agency’s own review criteria. The most studied single artifact is the Specific Aims page: a one-page document that, by near-unanimous practitioner consensus, determines a substantial portion of the application’s fate before reviewers read anything else. Empirical work on how reviewers allocate criterion scores (Eblen et al. 2016) supports this practitioner intuition: under the legacy five-criterion scoring, the Approach criterion and, to a lesser extent, the Significance criterion were the main predictors of the overall impact score and the funding outcome.

The 2024 transition to the NIH Simplified Review Framework restructured how these criteria are scored without removing them. Reviewers now assess three factors: Importance of the Research (Significance + Innovation, scored 1–9), Rigor and Feasibility (Approach, scored 1–9), and Expertise and Resources (Investigator + Environment, evaluated as sufficient/not sufficient rather than numerically scored). The five regulatory criteria persist; their packaging into factors changes the rhetorical work the Specific Aims page must do. NIH’s 2025 scoring guidance for the framework goes further than a repackaging: it tells reviewers that the Factor 1 score “should set a limit for the best possible overall impact score”, and that a strong approach cannot lift a project of moderate importance.1 Under that guidance, Importance becomes a ceiling and Approach can only hold or lower the score. The Eblen finding (Approach as the strongest predictor) comes from the legacy scoring regime and may not carry over unchanged.

Funding decisions have also changed downstream of review. Since the January 2026 Council round, NIH institutes no longer rely on paylines when deciding what to fund; they weigh peer-review information “in its entirety” alongside institute priorities and budgets.2 In August 2026 NIH asked for comment on a proposal to stop releasing numerical overall impact scores and instead report each application as “Most Competitive” (top 25%), “Competitive” (26–50%), or “Not Discussed”.3 NIH has also told applicants that applications “substantially developed by AI”, or containing sections substantially developed by AI, will not be considered the applicants’ original ideas.4

For scriptorium grant skills, this means three things. First, the Specific Aims page is the right unit of focus — it is structured, short, high-leverage, and follows a near-canonical four-paragraph shape. Second, “what makes a good aims page” is a mix of practitioner wisdom (well-documented across multiple workbooks but mostly non-empirical) and empirical findings (criterion-score correlations, percentile-funding relationships). The two should be labeled distinctly. Third, for NIH targets, any skill that produces aims-page prose must stay inside NIH’s originality policy. The declared-scaffolding scope already points in that direction, but the policy makes it binding.

Evidence and frameworks

Section titled “Evidence and frameworks”

Eblen et al. (2016), PLOS ONE 11(6):e0155060. Analyzed the relationship of the five criterion scores (Significance, Investigator, Innovation, Approach, Environment) to the Overall Impact score and to funding outcomes for over 123,700 competing R01 applications for fiscal years 2010–2013. The headline finding, in the abstract’s words: “the Approach and, to a lesser extent, the Significance criterion scores were the main predictors of an R01 application’s Overall Impact score and its likelihood of being funded.” Innovation, Investigator, and Environment were not named as main predictors. This corroborates the practitioner adage that a fundable aims page must convey not just that the science matters (Significance) but that the work can be done as proposed (Approach). The data predate the Simplified Review Framework.5

NIH Simplified Peer Review Framework (NOT-OD-24-010, effective January 2025). Reorganizes the five criteria into three factors:

  • Factor 1: Importance of the Research — combines Significance and Innovation. Scored 1–9.
  • Factor 2: Rigor and Feasibility — Approach. Scored 1–9.
  • Factor 3: Expertise and Resources — combines Investigator and Environment. Evaluated as sufficient / not sufficient.

The scale remains 1 (exceptional) to 9 (poor), with 5 as average. Overall Impact remains a separate 1–9 score that reviewers assign holistically.6 It applies to applications due on or after January 25, 2025 across most research project grant activity codes (R01, R03, R15, R21, U01, and others). NIH says it adopted the framework partly to address “the potential for reputational bias to affect peer review outcomes”, and that the binary Factor 3 rating “is intended to reduce the potential for general scientific reputation to have an undue influence.”6 The Significance criterion under the framework asks reviewers to “assess whether the application addresses an important gap in knowledge in the field, would solve a critical problem, or create a valuable conceptual or technical advance”, and to evaluate “the rigor of the scientific background for the work”. The Innovation criterion notes that “a project that is not applying novel concepts or approaches may be of critical importance for the field.”6

NIH Simplified Review Scoring Guidance (released March 14, 2025; effective October 2025 Council round). Tells reviewers how the factors combine into Overall Impact: “Logically, a project of moderate or limited importance (Factor 1) cannot be made more impactful by a strong approach (Factor 2) and/or appropriate expertise and resources (Factor 3)”; “Your Factor 1 score should set a limit for the best possible overall impact score”; and “Your assessment of Factors 2 and 3 can reinforce this score or worsen it.”1 NIH’s framework page also says that “an application does not need to be strong in all categories to be judged likely to have major scientific impact”.6 The two statements sit in some tension; the scoring guidance is the more specific instruction to reviewers.

NIH Specific Aims structural conventions. Practitioner literature converges on a four-paragraph aims page:

  1. Hook / significance opener. Establishes the problem and its scientific or clinical importance. Usually 3–5 sentences ending with the gap.
  2. Gap and rationale. Articulates what is not known, why the gap matters, and what the field’s working hypothesis is.
  3. Premise and central hypothesis. Names the long-term goal, the immediate objective, and the central hypothesis the proposal tests, anchored in the applicant’s own preliminary data.
  4. Aims block + impact close. Two or three specific aims (often bulleted), each a 2–4 sentence description of approach and expected result. Closes with a sentence on how the work advances the field.7

Practitioner sources (NIAID’s draft-specific-aims guidance, NINDS’s writing-specific-aims page, the Russell & Morrison Workbook) all prescribe variants of this structure. The convergence is striking and suggests that form has won: applicants who deviate substantially from this shape pay a legibility cost with reviewers who have internalized the convention.

Russell & Morrison, The Grant Application Writer’s Workbook (NIH version; Grant Writers’ Seminars & Workshops). The dominant practitioner workbook for NIH grant writing. Provides a step-by-step “bulleted outline” approach to the Specific Aims page: significance of the problem → gap in knowledge → long-term goal → objective → central hypothesis → rationale → specific aims → expected outcomes → impact. The workbook is regularly updated to track NIH form changes (current edition updated January 2026 for FORMS-I).8

Fang, Bowen & Casadevall (2016), eLife 5:e13323. “NIH peer review percentile scores are poorly predictive of grant productivity.” A re-analysis of a previously published dataset of NIH R01 grants funded 1980–2008, restricted to the 102,740 funded grants with percentile scores of 20 or better. Within that range, percentile rank had AUC = 0.54 (95% CI 0.53–0.54) for predicting above-median citation productivity — essentially equivalent to chance. The implication: within the funded range, peer review cannot reliably discriminate which grants will be most productive. The predictive validity of peer review for funding outcomes is real but constrained; once an application is good enough to be discussed, the noise floor is high.9

Eblen et al. extension on triage. Combining the above with the NIH triage process: in normal operation CSR states that “applications in the lower half are not typically discussed”, and streamlined applications receive no overall impact score.10 Among the discussed applications, percentile rank has historically predicted funding well at the percentile-payline boundary but predicts productivity weakly within the funded range. Practitioner implication: the aims page’s job is not to optimize for being “the best” — it is to get out of the streamlined bottom half with enough strength to be competitive. The older framing of this goal as landing “at or below the percentile payline” no longer describes NIH practice: institutes stopped relying on paylines from the January 2026 Council round,2 and the August 2026 proposal would replace released scores and percentiles with two discussed tiers.3

Emergency review modifications (January 2026 through January 2027 Council rounds). After the October–November 2025 lapse in appropriations, CSR changed how much of each meeting is discussed: “The top third (30 to 35 percent) will be discussed in the review meeting and considered for funding. The middle third will be designated as ‘competitive but not discussed’ and considered for funding. The lowest third will be designated as ‘not competitive and not discussed.’” Summary statements were shortened to a consensus sentence, bullets on the main score-driving points, and the assigned reviewers’ written critiques. NIH extended these modifications through the January 2027 Council round after receiving more than 40,000 applications for that round.11 So for applications reviewed in this period, the discussion line sits near the top third rather than the top half. Applications below it can still be funded if they fall in the “competitive” middle third. For those applications there is no discussion, so the written critiques carry most of the case (researcher inference; the shortened summary statement also has a consensus sentence and score-driving bullets).

NIH Unified Funding Strategy (announced November 21, 2025; effective January 2026 Council round). NIH states: “NIH ICOs will not rely on funding paylines in developing pay plans. Rather, ICOs will consider these scores in context of their and NIH’s priorities, strategic plans, and budgets.” Funding decisions “will still weigh peer review scores and critiques”, and institutes are to consider “peer review information in its entirety”, including career stage and the “broad distribution and geographic balance of funding”.2

Proposed change to reporting review outcomes (NOT-OD-26-088, request for information; comments closed October 13, 2026). NIH proposed that numerical overall impact scores “would no longer be released to principal investigators, applicant institutions, NIH program staff, ICO leadership, or NIH Advisory Councils”; applications would instead be placed in “Most Competitive” (top 25%), “Competitive” (26–50%), or “Not Discussed”. Written critiques, individual criterion and factor scores, and discussion summaries would continue to be released.3 As of this update it is a proposal, not adopted policy.

Centralized peer review. Effective with the August/October 2025 Council round, all NIH peer review of grants, cooperative agreements, and contracts is centralized within the Center for Scientific Review (CSR). NIH states that the review criteria, including the Simplified Review Framework, do not change.12

NIH policy on AI-developed applications (NOT-OD-25-132, effective September 25, 2025 receipt date). “NIH will not consider applications that are either substantially developed by AI, or contain sections substantially developed by AI, to be original ideas of applicants.” NIH says it uses AI-detection technology; if AI use is identified after award, NIH “may refer the matter to the Office of Research Integrity” and may disallow costs, suspend, or terminate the grant. The same notice limits each PI to six new, renewal, resubmission, or revision applications per calendar year (T activity codes and R13 conference grants excepted).4 The notice does not define “substantially”.

How this informs scriptorium

Section titled “How this informs scriptorium”

The aims page is a structured artifact, not free prose. A future specific-aims skill (planned for v0.4 in the [[DESIGN]] roadmap) should treat the aims page as a structured document with named slots: hook, gap, premise, hypothesis, aims-list, impact close. Critique and transformation skills on the aims page should target specific slots rather than the page as a whole.

Importance is a ceiling, not one factor among equals. Under the 2025 scoring guidance, a weak Factor 1 caps the overall impact score no matter how strong the approach.1 Aims-page critique for NIH targets should therefore check the importance slots (hook, gap, significance of the expected outcome) first, and should not tell an author that a strong approach section compensates for an unconvincing gap. This revises the earlier “Approach over Significance” emphasis drawn from Eblen et al.: that finding still describes the legacy regime, but for current applications the safer reading (researcher inference) is “establish Importance, then show feasibility.”

Criterion-aware critique. A grant-reviewer-simulation skill should be able to assume reviewer personas that map to the Simplified Framework factors: an “Importance of the Research” reviewer focused on whether Significance and Innovation are clearly articulated; a “Rigor and Feasibility” reviewer focused on Approach specifics, pitfalls, and alternative strategies; an “Expertise and Resources” reviewer who issues a sufficient/not-sufficient verdict. See reviewer-archetypes-grants for the persona structure.

No score-to-funding mapping. Skills must not translate a simulated or real score into a funding probability or a “below the payline” verdict. NIH no longer relies on paylines,2 and may stop releasing scores at all.3 Critiques should name the strengths and weaknesses reviewers are likely to write down, because those are the part of the review NIH says institutes will weigh. Under the emergency modifications, the written critiques are the main record for “competitive but not discussed” applications.11

Conservative-edit posture is even more important in grant contexts. Grant text is highly purpose-built: every sentence in the aims page is doing rhetorical work. The cost of an unauthorized rewrite is higher than for a manuscript discussion section, because the aims page is the document reviewers cite when arguing for or against funding. MANUSCRIPT_STATE.yaml for a grant project should default to strict preservation constraints; transformation skills should require explicit invocation.

NIH’s AI-originality policy bounds generation. NOT-OD-25-132 treats sections “substantially developed by AI” as not the applicant’s original ideas.4 A skill that renders aims-page prose for an NIH application is exposed to that policy even when the underlying aims were declared by the author. Grant skills targeting NIH should (a) prefer critique and slot-level suggestions the author writes in their own words over generated paragraphs; (b) state the policy in their output when target_funder is NIH; and (c) never present generated prose as ready to submit. This is researcher inference from the notice text; NIH has not said where its threshold lies.

Practitioner wisdom is labeled. Scriptorium should be clear in its outputs that structural recommendations on the aims page (the four-paragraph shape, the bulleted aims with expected outcomes, the impact close) come from practitioner consensus, not from peer-reviewed empirical evidence on what wins. The empirical evidence is on criterion-score correlations, not on rhetorical conventions.

Implementation priority for scriptorium

Section titled “Implementation priority for scriptorium”

Verdict: Yes (v0.4), with a changed scope for NIH targets.

If Yes: This note grounds the v0.4 specific-aims and aims-significance-coherence skills and the grant personas of reviewer-simulation. Phase: v0.4. Required data: project.target_type: grant, a target_funder declaration, the declared aims, hypotheses and significance, and the existing methods scaffolding.

Two changes since the May 2026 version of this note:

  1. aims-significance-coherence is the safer first deliverable (researcher inference). It is pure critique, so it cannot run into NOT-OD-25-132. It also targets the Factor 1 ceiling, the highest-leverage check under the 2025 scoring guidance.
  2. specific-aims should default to a slot audit plus suggestions for NIH targets, with prose rendering available only on explicit request and labeled with the NIH originality policy. Whether full prose rendering for NIH applications stays in scope is a roadmap decision to flag in the PR; this note does not make it.

Open questions / weak evidence

Section titled “Open questions / weak evidence”
  • The four-paragraph aims-page convention is well-documented in workbooks and institutional guidance but, as far as we have located, no large-N empirical study compares aims-page structure against funding outcomes. The convention may be a coordination equilibrium more than a causal advantage.
  • The relative weighting of the criteria varies by study section and by program officer culture. Generalizations across NIH institutes should be hedged.
  • The Simplified Framework is too recent (effective 2025) for outcome data; how the factor-based packaging changes review behavior is an open empirical question. CSR interviewed study-section chairs about review under the framework and published four video excerpts (September 2025; collected in an NIH Extramural Nexus post of January 14, 2026).13 As far as we located, NIH has published no quantitative analysis of factor scores against overall impact under the new guidance.
  • Whether Eblen et al.’s Approach-dominance survives the Factor 1 ceiling instruction is untested.
  • NOT-OD-26-088 is a proposal; whether NIH adopts score-free reporting, and in what form, was unknown at this update.
  • The emergency review modifications are scheduled to run through the January 2027 Council round. Whether the one-third discussion line and the “competitive but not discussed” category become permanent is unknown.
  • NOT-OD-25-132 does not define “substantially developed by AI”, and NIH has not published how its detection works. How the policy is applied in practice is unknown.

References

Section titled “References”

Footnotes

Section titled “Footnotes”

  1. NIH. Simplified Review Framework: Scoring Guidance. Released March 14, 2025; guidance effective 2025/10 Council round. https://grants.nih.gov/sites/default/files/Simplified-Review-Scoring-Guidance.pdf (accessed 2026-10-20). ↩ ↩2 ↩3

  2. NIH Extramural Nexus. Implementing a Unified NIH Funding Strategy to Guide Consistent and Clearer Award Decisions. November 21, 2025. https://grants.nih.gov/news-events/nih-extramural-nexus-news/2025/11/implementing-a-unified-nih-funding-strategy-to-guide-consistent-and-clearer-award-decisions ↩ ↩2 ↩3 ↩4

  3. NIH. NOT-OD-26-088: Request for Information (RFI) on Proposed Changes to Reporting Outcomes from NIH Peer Review. Comments due October 13, 2026. https://grants.nih.gov/grants/guide/notice-files/NOT-OD-26-088.html . See also NIH Extramural Nexus, “Your Feedback Wanted on Proposed Changes to Reporting Peer Review Outcomes for Applications”, August 18, 2026. https://grants.nih.gov/news-events/nih-extramural-nexus-news/2026/08/your-feedback-wanted-on-proposed-changes-to-reporting-peer-review-outcomes-for-applications ↩ ↩2 ↩3 ↩4

  4. NIH. NOT-OD-25-132: Supporting Fairness and Originality in NIH Research Applications. Released July 17, 2025; effective for applications submitted to the September 25, 2025 receipt date and beyond. https://grants.nih.gov/grants/guide/notice-files/NOT-OD-25-132.html ↩ ↩2 ↩3

  5. Eblen MK, Wagner RM, RoyChowdhury D, Patel KC, Pearson K. How criterion scores predict the Overall Impact score and funding outcomes for National Institutes of Health peer-reviewed applications. PLOS ONE. 2016;11(6):e0155060. doi:10.1371/journal.pone.0155060. PMID: 27249058. ↩

  6. NIH. NOT-OD-24-010: Simplified Review Framework for NIH Research Project Grant Applications. 2023. https://grants.nih.gov/grants/guide/notice-files/NOT-OD-24-010.html . NIH. Simplified Peer Review Framework (criteria text, activity codes, rationale). https://grants.nih.gov/policy-and-compliance/policy-topics/peer-review/simplifying-review/framework (accessed 2026-10-20). ↩ ↩2 ↩3 ↩4

  7. NIAID. Draft Specific Aims (practitioner guidance). https://www.niaid.nih.gov/grants-contracts/draft-specific-aims . NINDS. Writing Specific Aims. https://www.ninds.nih.gov/funding/preparing-your-application/preparing-research-plan/writing-specific-aims ↩

  8. Russell SW, Morrison DC. The Grant Application Writer’s Workbook: National Institutes of Health Version. Grant Writers’ Seminars & Workshops, LLC. Current edition: January 2026 (FORMS-I update). https://www.grantcentral.com/workbooks/national-institutes-of-health/ ↩

  9. Fang FC, Bowen A, Casadevall A. NIH peer review percentile scores are poorly predictive of grant productivity. eLife. 2016;5:e13323. doi:10.7554/eLife.13323. PMID: 26880623. ↩

  10. NIH Center for Scientific Review. What happens to your application during and after review? https://public.csr.nih.gov/ForApplicants/InitialReviewResultsAndAppeals/applicationduringafterreview (accessed 2026-10-20). ↩

  11. Reed B (CSR). Emergency Modifications to NIH Peer Review. NIH Extramural Nexus, November 24, 2025 (editor’s note 9/16/26). https://grants.nih.gov/news-events/nih-extramural-nexus-news/2025/11/emergency-modifications-to-nih-peer-review . NIH Extramural Nexus. Continuing Modified Peer Review Processes for the January 2027 Council Round. September 18, 2026 (NOT-OD-26-012, NOT-OD-26-069, NOT-OD-26-114). https://grants.nih.gov/news-events/nih-extramural-nexus-news/2026/09/continuing-modified-peer-review-processes-for-the-january-2027-council-round ↩ ↩2

  12. NIH. NOT-OD-25-135: Notice to Applicants of Adjustments to Achieve Consistency in Peer Review Contacts and Processes. 2025. https://grants.nih.gov/grants/guide/notice-files/NOT-OD-25-135.html ↩

  13. NIH Extramural Nexus. Learn What Reviewers Say About Implementing the Simplified Review Framework. January 14, 2026. https://grants.nih.gov/news-events/nih-extramural-nexus-news/2026/01/learn-what-reviewers-say-about-implementing-the-simplified-review-framework ↩