---
name: research-proposals
description: Develop a research or funding proposal with a verified problem statement, coherent aims, feasible methods, risks, milestones, and evidence-bounded significance.
---

# Research proposals

Use this Academic Research Assistant skill for thesis, project, fellowship, and grant proposals. Apply the [research contract](../../docs/RESEARCH-CONTRACT.md). A proposal describes intended work and the evidence that makes it worth attempting. Keep preliminary findings, established results, and future ambitions clearly distinguishable.

## Inputs and constraints

Obtain the call or institutional requirements, audience, evaluation criteria, topic, project duration, resources, team roles, existing evidence, and required sections or limits. Verify current requirements from the actual call when available. Do not invent eligibility, budget rates, institutional commitments, collaborators, approvals, or facility access.

Identify the proposal stage: concept note, full application, or revision after feedback. Preserve the user's scientific ownership and existing commitments. A missing detail that affects feasibility belongs in an open-decisions list; routine narrative choices can proceed with stated assumptions.

## Construct the argument

1. Write the problem in terms of a specific uncertainty or limitation and its consequence. Support the current state of knowledge with verified sources. Distinguish “important to study” from “nobody has studied this,” which requires a defensible search and remains bounded by its coverage.
2. Define an overall objective and a small set of coherent aims. Each aim should have a question, evidence or reasoning strategy, expected output, and a criterion for interpreting its outcome. Avoid making every aim depend on an untested success in the first aim without a contingency.
3. Link the proposed contribution to the limitation. Explain how the approach could change knowledge, capability, or practice, and identify alternative explanations or approaches. Do not promise an unprecedented result, guaranteed effect, publication, or downstream impact.
4. Build the research plan through [methodology design](../methodology-design/SKILL.md). Specify selection or domain, procedures, analysis, validation, data management, and ethics needs at a depth appropriate to the application. Methods should answer the aims rather than function as an equipment inventory.
5. Establish feasibility from actual preliminary work, expertise, access, and resources. Label pilot evidence with its size, scope, and unresolved limitations. A collaborator's proposed role or an application for access is not confirmed support.
6. Define milestones, dependencies, decision points, and fallback paths through [research planning](../research-planning/SKILL.md). Include informative outcomes when the main hypothesis fails. Do not equate a calendar deadline with evidence that the work has succeeded.
7. Align costs and effort with activities. Use supplied or verified rates and mark estimates. Explain major resource needs; do not fabricate quotations, salaries, overhead policies, or institutional approvals to complete a budget.

## Adapt to the research mode

Quantitative proposals need a defensible target quantity, measurement plan, design assumptions, uncertainty strategy, and sample-size or precision rationale. State assumptions behind power calculations and alternatives if recruitment or data quality falls short.

Qualitative proposals need a coherent methodological stance, purposeful access or sampling, data-generation plan, analytic strategy, reflexive considerations, and a reasoned account of adequacy. Avoid promises of statistically representative conclusions unless a suitable sampling design supports them.

Theoretical proposals should name the mathematical objects or conceptual premises, proposed derivations or arguments, limiting cases, and intermediate results that make progress assessable. Computational proposals should separate formulation, implementation, numerical verification, benchmarking, and scientific validation, with realistic resource and reproducibility plans.

## Deliverable and verification

Deliver an editable proposal shaped to the actual requirements, an aims-to-methods-to-outputs map, a milestone and risk table, references, and a clearly marked list of unresolved commitments. Use [manuscript writing](../manuscript-writing/SKILL.md) only for research narrative components that need a manuscript-style treatment; preserve the proposal's future tense and assessment purpose.

Check compliance with section limits, internal consistency, arithmetic, roles, dates, and source support. Verify that every aim has an executable method and every claimed capability has evidence. State whether the result is a concept draft, a complete application draft, or a checked submission package. Preparing documents does not authorize external submission or certify institutional approval.