This module develops proposal writing, criterion-based review, and responsible use of AI in proposal work. Students learn to identify the governing funding opportunity, make evidence-calibrated claims, write and revise a research narrative, and review proposals without substituting prestige, unsupported assumptions, or tool output for scientific judgment.
Planning context: A prior anonymous class survey—the cohort year is not documented in the repository—had varying response counts across items (approximately 14–18). Respondents most often identified novelty, clarity, feasibility, background/significance, research design/methods, and aims/hypotheses as concerns. These small, item-specific results guide practice time; they are not treated as representative evidence about graduate students generally.
Teaching pattern: Each 80-minute meeting has a bounded source or anchor task, a short application or decision task, a concrete output, and closure time. Historical examples are used to inspect evidence and framing—not to infer universal causes of scientific or funding success.
| # | Lecture | Key Topics and Output |
|---|---|---|
| 1 | Research Pathways & Evidence-Consistent Framing | Source-bound landmark cases; classification limits; paragraph reframe |
| 2 | Funding Agencies & Your First Research Narrative | Current NSF/NIH review frameworks; challenge/objectives draft |
| 3 | Writing the Research Narrative | NIH Specific Aims as an argument; revision and elevator test |
| 4 | Criteria Calibration: Intellectual Merit, Broader Impacts & Evidence | Synthetic proposal comparison; evidence-based calibration memo |
| 5 | AI in Proposal Development: Use, Evidence & Confidentiality | Role-specific policy; safe workflow; AI-use and verification note |
| 6 | Formal Peer Review & Responding to Critique | Individual review; panel summary; response-to-critique memo |
| 7 | GCR-Inspired Team Concept Workshop | Convergence concept note; cross-team review and revision plan |
Individual: Research Narrative Draft
| Lecture | Milestone |
|---|---|
| 2 | Draft challenge statement and a bounded set of research objectives |
| 2 | Cross-disciplinary criterion-based peer review |
| 3 | Revised challenge/objectives and 90-second elevator test |
| 5 | AI-use and verification note for any AI-assisted work; submit “no AI used” when applicable |
Students may work on a dissertation or qualifying-exam topic only if it is appropriate to share with classmates. A public, published, or instructor-created alternative is always available; nobody is required to disclose confidential or unpublished research.
Team: GCR-Inspired Concept Note
| Lecture | Milestone |
|---|---|
| 4 | Select the problem and necessary disciplinary areas; draft an integration map |
| 5 | Add a risk/feasibility note and AI-use/verification protocol |
| Pre-L7 | Upload a complete two-page draft before class |
| 7 | Cross-team review, reviewer–author conference, revision, and response-to-review note |
| Post-L7 | Final revised concept note due on Blackboard October 14 |
This is a short course exercise modeled on an authentic NSF genre. It is not represented as a complete or submission-ready NSF GCR proposal.
Goal: Distinguish observation-, hypothesis-, and method-centered pathways without forcing complex research into a single category, then reframe a paragraph for a target audience without changing the underlying research or evidence.
For each source, students mark: the event described, the evidence passage supporting a classification, one plausible alternative classification, and one fact the source does not establish.
Interpretive limit: “Hypothesis-centered,” “observation-centered,” and “method-centered” are course lenses, not Nobel Committee labels. Prize records also foreground the award and laureates; they are not complete histories of every contributor or cause. This three-case packet cannot support percentages or trends across disciplines.
| Time | Activity |
|---|---|
| 8 min | Frame the source rule: separate text-supported fact, interpretation, and missing evidence |
| 18 min | Triads classify the three landmark cases and cite the passage supporting each classification |
| 12 min | Compare classifications; identify mixed pathways and what the sources leave unresolved |
| 22 min | Reframe a public, published, or instructor-created research paragraph for a specified funding audience without changing the proposed work |
| 12 min | Peer audit: underline changed claims, unsupported predictions, hidden audience assumptions, and retained applied context |
| 5 min | Synthesis: when audience adaptation is accurate and when it becomes spin |
| 3 min | Exit ticket: one defensible reframe plus one evidence boundary |
| 80 min | Total |
Goal: Read a funding opportunity and current review framework before drafting a challenge statement and objectives that fit the audience without treating a course template as a sponsor rule.
The instructor supplies one bounded NSF solicitation excerpt and one bounded NIH NOFO excerpt. Before drafting, each student records the agency, program or mechanism, document number/version and date, relevant review criteria, and any instruction that changes the general comparison below. Students may select one of these fictionalized/public opportunities as the audience for their draft.
| Feature | NSF | NIH |
|---|---|---|
| Governing instruction | Current PAPPG, supplements, and the specific solicitation | Current application guide and the specific notice of funding opportunity (NOFO) |
| Core review framework | Intellectual Merit and Broader Impacts; solicitations may add criteria | For most covered research project grants with due dates on or after January 25, 2025, three factors organize five regulatory criteria: Importance of the Research; Rigor and Feasibility; Expertise and Resources |
| Assessment | Reviewers apply the criteria and relevant solicitation; rating and panel practices can vary by competition | Factors 1 and 2 are scored 1–9; Factor 3 is evaluated as “Appropriate” or “Gaps Identified”; all inform Overall Impact |
| Typical space | Many PAPPG proposals allow a 15-page Project Description, but the solicitation controls | Specific Aims is generally 1 page and an R01 Research Strategy generally 12 pages, but the NOFO and activity code control |
| Preliminary evidence | No agency-wide preliminary-data requirement; evidence should support the rationale and feasibility expected by the solicitation and proposed work | New applications should present pertinent studies, data, and/or experience. For many mechanisms, preliminary evidence can be important to feasibility; expectations depend on the NOFO and project |
| Societal outcomes | Intellectual Merit and Broader Impacts both receive full consideration; each is necessary and neither is sufficient by itself | Broader Impacts is not a separate scored factor; public-health relevance and other requirements are addressed through the applicable framework and application sections |
| Framing | Fit the program and solicitation while representing the work accurately | Fit the institute/center, NOFO, activity code, and review audience while representing the work accurately |
Policy status—checked August 6, 2026: NSF 24-1 and Supplements NSF 26-200 and NSF 26-202 remain the current PAPPG materials. NSF has published a draft Guidance on Financial Assistance intended to replace the PAPPG; comments close August 24, 2026, but the draft is not current policy. Recheck the maintained landing page before this lecture and before any real submission.
| Time | Activity |
|---|---|
| 15 min | Opportunity check and source-guided comparison: record the assigned document, criteria, score language, and opportunity-specific exceptions |
| 18 min | Draft a 4–6 sentence research challenge |
| 12 min | Draft two or three research objectives with a clear relationship to the challenge |
| 15 min | Cross-disciplinary peer review against the stated audience and relevant criteria |
| 12 min | Revise the challenge and objectives from the review |
| 5 min | Source check: label one sentence as supported, qualified, or still needing evidence |
| 3 min | Exit ticket and submission |
| 80 min | Total |
Students may use this challenge scaffold:
While [current capability or importance], a major unresolved problem is [bounded knowledge or capability gap], which limits [specific consequence]. Existing work does not yet establish [what is missing and why].
Students may also test a foundational → mechanistic → application/validation objective sequence. These are optional course tools, not NSF or NIH requirements. Use a different sequence when the science demands it, and do not imply that later work is feasible merely because it appears third in a list.
The Lecture 2 draft is formative or lightly weighted. Lecture 3 assesses the documented revision rather than charging students twice for the same text within 48 hours.
Output: Challenge/Objectives Draft submitted on Blackboard.
Goal: Analyze how a one-page Specific Aims document functions as an argument, then revise the Lecture 2 narrative for logic, evidence, feasibility, and a non-specialist reader.
The instructor selects one complete aims page from the NIAID Sample Applications and Documents collection and preserves its wording and attribution under the permission stated with that sample. Students also use the current NIH Research Plan instructions and the target NOFO.
Historical limit: A posted sample can illustrate grantsmanship, but it may reflect an older form, review framework, or funding context. Its award does not prove that a particular sentence or structure caused success. Current instructions and the target NOFO control.
Students annotate the page for a commonly used argument sequence:
This sequence is descriptive guidance, not a mandatory paragraph formula.
| Time | Activity |
|---|---|
| 10 min | Compare the purpose and audience of an NSF narrative and an NIH Specific Aims page |
| 15 min | Individually annotate the public NIH aims page for claims, evidence, gap, approach, outcomes, and uncertainty |
| 10 min | Build an argument map; compare interpretations and identify what the page alone cannot establish |
| 18 min | Revise the Lecture 2 challenge for specificity, evidence, and consequence |
| 12 min | Stress-test objective dependencies, success indicators, alternatives, and feasible scope |
| 10 min | Paired 90-second elevator test: “What will you learn?” and “Why does it matter?” |
| 5 min | Final revision note and submission |
| 80 min | Total |
Students using sensitive work may complete the revision with an instructor-created composite. Peer partners respond to the argument presented; they are not asked to judge undisclosed proprietary details.
Output: Revised Challenge/Objectives plus a two-sentence revision note.
Goal: Apply current NSF criteria consistently, distinguish evidence from inference, and identify when a review substitutes reputation or a generic rule for the proposal record.
NSF requires full consideration of both Intellectual Merit and Broader Impacts; this is not a universal numerical 50/50 weighting rule. Solicitations may add criteria, and funding decisions can also reflect portfolio considerations beyond a classroom rating exercise.
Both excerpts are instructor-created composites, not real proposals or records of actual funding decisions.
Excerpt A—novel integration, bounded evidence: An early-career team proposes a bioinspired soft-sensor architecture to test whether a distributed structural mechanism can separate deformation modes that conventional sensors conflate. Component calibration and noise data are available, but an integrated prototype has not been demonstrated. The plan specifies controls, decision milestones, an alternative architecture, needed facilities, and complementary team expertise. Its Broader Impacts plan co-designs an open laboratory module with two community-college instructors and defines adoption and learning measures.
Excerpt B—extensive data, unclear advance: An established team proposes to optimize an existing sensor-analysis pipeline. Pilot data show a modest accuracy gain on the team’s prior dataset, but the excerpt does not state what generalizable knowledge would result or whether the gain persists outside that dataset. Its Broader Impacts paragraph promises a summer workshop for “diverse students” but identifies no partner, participant need, access plan, activity, or measure of effect.
The purpose is not to identify a hidden winner. Students identify supported strengths, material uncertainties, and the additional evidence needed under the stated solicitation.
| Time | Activity |
|---|---|
| 12 min | Source briefing: criteria, review elements, solicitation-specific additions, and limits of classroom ratings |
| 8 min | Individual criterion-linked notes on Excerpts A and B |
| 12 min | Calibration triads compare ratings and mark where evidence, inference, or an unstated preference drove disagreement |
| 8 min | Build a claim–evidence–uncertainty map for one excerpt |
| 10 min | Module 2 Microcase 2: “No Famous Name, No Preliminary Data”—use the NSF branch: 8-minute case plus 2-minute debrief |
| 12 min | Redesign one Broader Impacts paragraph: audience/partner, need, activity, responsible actor, access, and assessment |
| 5 min | Write an individual calibration memo: one strength, one bounded gap, and one conditional request |
| 10 min | Launch the GCR-inspired concept: shortlist a problem and map which disciplinary areas must integrate |
| 3 min | Exit ticket: one rating assumption to monitor in Lecture 6 and the team’s next concept task |
| 80 min | Total |
Outputs: Each student retains a short three-part calibration memo for the formal panel. Each team posts its initial GCR integration map in the course workspace as a concept-note milestone.
Goal: Distinguish applicant, reviewer, sponsor, course, and tool rules; classify proposal information before tool use; verify AI-mediated claims; and document human responsibility.
| Role | Rule to Apply |
|---|---|
| NIH reviewer | Generative AI may not be used to analyze applications or formulate critiques; uploading application, proposal, critique, or original-concept content to an online generative-AI tool violates NIH confidentiality requirements |
| NSF reviewer | Proposal and review information may not be uploaded to non-approved generative-AI tools; NSF separately permits sharing publicly available information with generative-AI tools, subject to other applicable rules |
| NIH applicant | Limited assistance may be possible, but NIH does not treat applications or sections substantially developed by AI as the applicant’s original ideas; applicants remain responsible for integrity and compliance |
| NSF proposer | NSF encourages disclosure of generative-AI use and holds proposers responsible for the proposal’s accuracy and authenticity |
| Student in this course | Follow the course AI Use Policy, disclose AI-generated content used in an assignment, verify original sources, and also follow sponsor, institutional, data-use, privacy, IP, and tool rules |
Course permission never overrides a sponsor’s rule or authorizes disclosure of unpublished, proprietary, privileged, personally identifiable, or review-protected material.
| Time | Activity |
|---|---|
| 8 min | Retrieve Module 1 norms: disclosure, verification labor, input classification, and sign-off |
| 15 min | Policy jigsaw: NIH reviewer, NIH applicant, NSF reviewer/proposer, and course/tool branch |
| 10 min | Role comparison: what changes when the same person moves from applicant to reviewer to student? |
| 12 min | Module 2 Microcase 1: “The Convenient Review Assistant”—8-minute case plus 4-minute policy debrief |
| 15 min | Verify three constructed claims: “NSF uses a fixed 50/50 weighting”; “NIH always requires preliminary data”; “disclosure permits AI to draft any proposal section” |
| 12 min | Build a safe workflow and AI-use/verification note; teams add a risk/feasibility note and protocol to the GCR concept |
| 5 min | Pair check for hidden protected inputs, missing verification, or unclear human sign-off |
| 3 min | Exit ticket: permitted task, prohibited input, responsible verifier |
| 80 min | Total |
Record the task; tool and version when available; date; input classification; output retained or used; claims and original sources checked; material changes; human verifier/signatory; required disclosure; and any unresolved limitation. Students who used no AI submit “No AI used.”
Documenting any AI assistance is a Module 2 assignment requirement. It is not presented as a universal NSF, NIH, institutional, or journal disclosure standard.
Individual product: Submit the AI-use/verification note with the individual narrative record on Blackboard; submit “No AI used” when applicable. Team product: Add the risk/feasibility note and AI-use/provenance protocol to the working GCR-inspired concept draft; these are assessed with the final October 14 concept note rather than as a second stand-alone submission.
This class does not ask students to upload any proposal to an AI tool. Exercises use public, synthetic, or explicitly authorized text, and policy analysis can be completed without using AI.
Module 2 addresses proposal-specific permission, confidentiality, accuracy, and workflow. Module 3 evaluates AI feedback and source verification in scientific writing. Module 4 synthesizes broader ethical reasoning, governance, accountability, and research-lifecycle risks. A generic AI-history quiz or detection contest is not needed here.
Goal: Produce an independent criterion-based review, deliberate without losing minority concerns, write an accurate panel summary, and convert critique into a bounded revision response.
The instructor supplies a two-page synthetic or explicitly authorized proposal excerpt, its fictional funding-opportunity excerpt, the current review framework, and role cards. The packet must be different from Lecture 4’s calibration excerpts and must contain enough information to identify evidence, uncertainty, and at least one genuine tradeoff. It must not include a student’s unpublished proposal without specific permission and an equivalent alternative.
Before reading, students declare a fictional conflict from their role card or “no conflict,” identify the packet’s confidentiality classification, and agree that criteria—not seniority, eloquence, or name recognition—govern the discussion. No generative-AI review is used in this simulation.
| Time | Activity |
|---|---|
| 10 min | Reviewer briefing: role, conflicts, confidentiality, framework, evidence language, and chair/recorder duties |
| 15 min | Independent written review before discussion |
| 20 min | Panel deliberation; the chair records agreement, disagreement, open questions, and criterion-linked reasons |
| 15 min | Draft a panel summary that does not erase a material minority concern |
| 12 min | Author perspective: write a response-to-critique action, clarification, or reasoned disagreement |
| 5 min | Fairness and power debrief using the Lecture 4 assumption each student chose to monitor |
| 3 min | Exit ticket: one panel-process safeguard |
| 80 min | Total |
Output: Team Panel Summary submitted on Blackboard; individual response-to-critique memo retained in the student’s course portfolio.
Goal: Integrate proposal framing, review criteria, and Module 1 team procedures in a short concept note addressing a problem that requires deep disciplinary integration.
Status checked August 6, 2026: GCR is active. Students must use the current program page, and an instructor must recheck status before class. The course assignment is a GCR-inspired concept note, not a full GCR proposal: it does not reproduce the solicitation’s complete scale, phases, research plan, management plan, budget, or submission requirements.
“While…However” and the three-objective sequence may be used as course writing scaffolds. Neither is an official GCR requirement.
| Pattern | Diagnostic Question |
|---|---|
| Parallel contribution | Could each discipline complete its piece with only a handoff? |
| Interdisciplinary integration | Are methods or concepts combined to answer a shared question? |
| Convergence aspiration | Does sustained integration change how participants frame the problem or create capacity none could produce alone? |
These are course diagnostics, not official categorical scores. Teams should explain the integration mechanism rather than merely select a label.
| Time | Activity |
|---|---|
| 8 min | Scope the authentic solicitation and the limits of the course exercise |
| 10 min | Paired pitch and convergence test: problem, necessary disciplines, handoffs, and reciprocal integration |
| 15 min | Silent cross-team criterion-based review of the complete pre-class draft |
| 10 min | Reviewer–author conference focused on evidence and one question that could change the review |
| 22 min | Team revision sprint on the problem, integration, objectives, IM/BI, fit, and management plan |
| 10 min | Record accepted, modified, and declined feedback with reasons; complete final compliance check |
| 5 min | Exit ticket: remaining revision, owner, deadline, and evidence needed |
| 80 min | Total |
This is an instructor-created course rubric, not NSF’s official GCR review or scoring scale. Rate each item 1–5 and cite one passage:
Add one prioritized revision and one question whose answer could change the review.
Output: Review, response note, and revisions completed in class; revised final concept note due October 14.
| Material | Status and Required Boundary |
|---|---|
| Lecture 1 Nobel packet | Ready from the linked primary prize records; teach selected-case and incomplete-history limits |
| Lecture 3 NIH aims page | Select from the NIAID public sample collection, preserve attribution/permission terms, label its date, and pair it with current instructions |
| Lecture 4 calibration excerpts | Ready as constructed cases on this page; do not present them as real awards |
| Lecture 5 policy claims and microcase | Ready from current primary policy sources; recheck before teaching |
| Lecture 6 formal panel packet | Instructor preparation still required: synthetic or authorized excerpt, fictional opportunity, current rubric, role/conflict cards, and accessibility review |
| Lecture 7 concept-note instructions | Ready as a course exercise; recheck the active GCR page before teaching |
The repository retains earlier presentation files, but they are not assigned as evidence in this revision. Before reuse:
Until those checks are complete, use the source-bound and constructed activities above. Do not infer that an award outcome reveals which revision caused funding or that an AI comparison without a reproducible packet demonstrates a general model limitation.
Before the relevant class, the instructor should:
By the end of this module, students should be able to:
Detailed submission portals and any instructor-supplied packets are available on Blackboard. Public instructions on this page define the course expectations and evidence boundaries.