Grant Writing

---
name: grant-writing
description: "Translate research vision into funded reality."
lastReviewed: 2026-04-30
---

# Grant Writing Skill


> Translate research vision into funded reality.

## Core Principle

A grant proposal is a persuasion document. You're convincing reviewers that your research is important, your approach is sound, and you're the right person to do it.

## The Three Questions Every Reviewer Asks

1. **Why does this matter?** (Significance)
2. **Will this approach work?** (Feasibility)
3. **Can this team do it?** (Expertise)

Answer all three. Clearly. Early.

## Proposal Structure (General)

### 1. Specific Aims (1 page)

The most important page. Reviewers decide here.

**Structure:**

- **Hook**: Why this problem matters NOW
- **Gap**: What's missing in current knowledge
- **Hypothesis**: Your central claim
- **Aims**: 2-4 specific, achievable objectives
- **Impact**: What changes if you succeed

**Template:**
> [Problem statement establishing significance]. Despite [current state of knowledge], [the gap] remains unaddressed. We hypothesize that [your central hypothesis]. To test this, we will: Aim 1: [specific objective]. Aim 2: [specific objective]. This work will [impact statement].

### 2. Significance & Innovation

- Why the problem matters (societal, scientific, economic)
- What's new about your approach
- How it advances the field

### 3. Approach / Research Plan

- Detailed methodology for each aim
- Preliminary data (shows feasibility)
- Timeline and milestones
- Potential pitfalls and alternatives

### 4. Investigator Qualifications

- Why you're the right person/team
- Relevant expertise and publications
- Collaborations and resources

### 5. Budget & Justification

- Personnel, equipment, supplies, travel
- Clear justification for each item
- Matches scope of work

```yaml
# Sample 3-year budget (NSF format)
budget:
  year_1:
    personnel:
      - role: "PI (2 months summer)"
        amount: 24000
      - role: "Graduate RA (12 months)"
        amount: 35000
    fringe_benefits: 17700  # 30% of salaries
    equipment: 15000        # Specialized hardware
    supplies: 5000
    travel: 3000            # 1 conference
    indirect: 42280         # 52% MTDC
    total: 141980
  total_request: 425940

justification: |
  PI will supervise graduate student and conduct analysis.
  Equipment: Neural network training workstation essential for Aim 2.
  Travel: Present findings at ACM CHI.
```

## Agency-Specific Guidance

### NSF (National Science Foundation)

- **Broader Impacts** required (education, diversity, public benefit)
- **Intellectual Merit** equally weighted
- Project descriptions limited to 15 pages
- Annual reports and data management plan required

### NIH (National Institutes of Health)

- **Significance, Innovation, Approach, Investigators, Environment** (5 criteria)
- R01 is the standard research grant
- K awards for career development
- Page limits vary by mechanism
- Biosketch format is strict

### Private Foundations

- Often shorter applications
- More flexibility in format
- Relationship building matters
- May prefer specific populations or approaches

## Writing Strategies

### The Inverted Pyramid

Start with the most important information:

1. **Significance** (why care?)
2. **Innovation** (what's new?)
3. **Approach** (how?)
4. **Details** (specifics)

### Active Voice, Concrete Claims

| Weak | Strong |
|------|--------|
| "It is believed that..." | "We will test whether..." |
| "Studies will be performed" | "We will conduct experiments" |
| "This may lead to..." | "This will demonstrate..." |

### Preliminary Data Strategy

- Show you CAN do the work
- Demonstrate feasibility, not completion
- Just enough to prove concept
- Save some results for the funded project

### Addressing Weaknesses

- Acknowledge risks upfront
- Provide alternatives for each
- Shows you've thought it through
- Reviewers find problems anyway—beat them to it

## Common Mistakes

| Mistake | Fix |
|---------|-----|
| Burying the significance | Lead with impact |
| Too much jargon | Write for educated non-expert |
| Vague aims | Make aims specific and measurable |
| No preliminary data | Pilot studies, even small ones |
| Ignoring page limits | Ruthless editing |
| No alternatives | "If X fails, we will Y" |
| Weak budget justification | Every dollar explained |
| Missing required sections | Use the checklist |

## Review Criteria Alignment

Map your writing to review criteria:

| Criterion | Where to Address |
|-----------|------------------|
| Significance | Specific Aims, Significance section |
| Innovation | Specific Aims, Innovation section |
| Approach | Research Plan, each aim |
| Investigator | Biosketch, Team section |
| Environment | Resources, Letters of support |

## The Review Process (Know Your Audience)

1. **Assignment**: Program officer assigns to study section
2. **Primary reviewers**: 2-3 read in detail, score each criterion
3. **Panel discussion**: Top 50% discussed
4. **Scoring**: 1 (best) to 9 (worst) for each criterion
5. **Funding line**: Percentile determines funding

**Key insight**: Reviewers are tired, busy experts. Make it EASY to find your strengths.

## Timeline for Submission

| Weeks Before | Task |
|--------------|------|
| 12+ | Start Specific Aims draft |
| 10 | Circulate Aims for feedback |
| 8 | First draft of full proposal |
| 6 | Internal review |
| 4 | Major revisions complete |
| 2 | Final polish, budget finalized |
| 1 | Institutional review |
| 0 | Submit (never day-of!) |

## Resubmission Strategy

Most grants don't fund on first try. Resubmissions:

- Address EVERY reviewer concern
- Show what changed (clearly marked)
- Don't argue with reviewers—adapt
- Include new preliminary data
- Resubmit to same study section if possible

## Budget Tips

- **Modular budgets** ($250K/year blocks) for NIH R01
- **Match effort to work** (if you're doing half the work, request half time)
- **Justify everything** (why this equipment? why this travel?)
- **Include indirect costs** (check your institution's rate)
- **Don't under-budget** (reviewers wonder what you're hiding)