Scholarship SignalScholarships & research funding
AccountGet started
Application Intelligence14 min read

How to Write a Research Proposal for Scholarship Applications

Move from a broad topic to a feasible question, evidence base, method and expected contribution that fits the funding period.

Scholarship Signal editorial team, Application guidanceReviewed by Scholarship Signal verification deskPublished 14 March 2026Updated 27 July 2026
14 minReading time
27 July 2026Last reviewed update
Source reviewedOfficial rules stay authoritative
In this guide

The short answer

A research proposal must show that your question is worth answering, that you know how to answer it, and that it can be answered within the funding period with the resources available. Feasibility is assessed as heavily as originality, and a narrower question with a credible method beats a broad one without.

Define a researchable problem

A strong proposal identifies a specific gap or unresolved decision and shows why it matters. The problem must be narrow enough to investigate with the available time, data and methods, which usually means narrower than feels comfortable.

The most common failure is a topic rather than a question. 'Climate adaptation in coastal cities' is a topic. 'Whether municipal flood-insurance subsidies changed rebuilding location choices in three Gulf Coast cities after 2017' is a question.

Proposal structure

  1. State the problem and the central research question in the first paragraph.
  2. Summarise the most relevant literature and identify the specific gap, not the whole field.
  3. Define the conceptual or theoretical framing you are working within.
  4. Describe data, sample, variables, instruments and methods concretely.
  5. Address feasibility, access, ethics and limitations honestly.
  6. Explain the expected academic and practical contributions.
  7. Provide a realistic work plan mapped to the funding period, and references.

Method alignment

  • Each method answers a stated research question, and each question has a method.
  • The required data can realistically be obtained, and you say how.
  • The unit of analysis, sample frame and time period are explicit.
  • Validation, robustness or triangulation checks are described.
  • Ethical approval and data-protection requirements are identified with a route to obtaining them.

Make feasibility visible

Reviewers reject proposals that could not be completed in the funded period more often than they reject unoriginal ones. A three-year studentship that requires four years of fieldwork is a rejection regardless of the idea's quality.

Include a work plan by phase with the months allocated, and name what you will drop if access or recruitment is slower than planned. Acknowledging a risk with a mitigation reads as competence, not weakness.

Before submission

  • The proposal follows the funder's exact headings, length and reference style.
  • Terminology is consistent throughout, especially for key constructs.
  • Every claim about the literature is supported by a cited source.
  • The plan is feasible within the scholarship period, with a stated contingency.
  • A non-specialist in your department can state your question after one read.

Example: Research question

Weak version

This research will explore the impact of social media on young people's mental health, which is an important and under-researched area with significant implications for society.

Stronger version

This project asks whether the 2024 removal of algorithmic feed ranking for under-18 accounts in the EU changed self-reported anxiety among 13–17 year olds, using the difference between EU and UK cohorts in the existing Millennium Cohort and HBSC panels as the identifying variation.

Why the second version works

  • It names a specific, dated intervention rather than a broad phenomenon.
  • It identifies the population, the outcome and the data sources.
  • It states the identification strategy, so a reviewer can judge whether the design can support a causal claim.
  • 'Under-researched' is dropped: it is asserted in almost every weak proposal and evidenced in few.

Example: Feasibility paragraph

Weak version

I will conduct interviews with participants and analyse the data using appropriate qualitative methods. I am confident the research can be completed within the funding period.

Stronger version

Recruitment targets 24 clinic managers across four districts, drawn from the provincial registry I already have access to through my current role. Based on a 40% response rate in a comparable 2025 survey, this requires approaching approximately 60. Interviews run in months 6–13, transcription and coding in 10–18, leaving months 19–24 for writing. If recruitment falls below 18 by month 10, I will narrow to two districts and increase depth rather than extend fieldwork.

Why the second version works

  • Sample size is justified with a comparable response rate, not asserted.
  • Access is evidenced by an existing relationship rather than assumed.
  • The timeline is mapped to specific months within the funded period.
  • The contingency shows the project degrades gracefully instead of failing.

Research proposal skeleton

Where the funder supplies headings, use theirs. This structure is for applications that specify only a length.

TITLE
[Specific enough to convey the question, not just the topic.]

1. PROBLEM AND RESEARCH QUESTION (approx. 15%)
[The unresolved problem.] [Why it matters, with evidence.]
Central question: [One sentence.]
Sub-questions: [Two to four, each answerable by a stated method.]

2. LITERATURE AND GAP (approx. 20%)
[The two or three strands of work your question sits between.]
[The specific gap — not "little research exists" but what is missing and why it matters.]

3. CONCEPTUAL FRAMING (approx. 10%)
[Theory or framework, and the constructs it commits you to.]

4. DATA AND METHODS (approx. 30%)
Design: [Approach and why it fits the question.]
Data: [Source, access route, unit of analysis, time period, sample size and justification.]
Analysis: [Specific techniques, per sub-question.]
Validation: [Robustness, triangulation or reliability checks.]

5. FEASIBILITY, ETHICS AND RISK (approx. 10%)
[Access.] [Ethical approval route.] [Data protection.]
[Main risk and the mitigation.]

6. CONTRIBUTION (approx. 10%)
[Academic contribution.] [Practical or policy contribution.]

7. WORK PLAN
Months 1-6: [...]   Months 7-12: [...]   Months 13-24: [...]   Months 25-36: [...]

REFERENCES
[Funder's required style.]

Common reasons applications are rejected

  • A topic instead of an answerable research question.
  • A scope that cannot be completed within the funding period.
  • No stated route to the data, or an assumption of access that is not evidenced.
  • Methods that do not match the questions, or questions with no method attached.
  • A literature review that summarises the field instead of locating a specific gap.
  • Ethics treated as a formality when the population or data is sensitive.
  • Ignoring the funder's required headings, length or reference style.
  • No contribution stated beyond adding to the literature.

How this changes by scholarship type

Doctoral studentships attached to a project
The topic is largely fixed. Your proposal shows how you would approach the advertised project, so engage with the named supervisor's recent work directly.
Open doctoral competitions
You choose the question, so fit with the department's existing strengths is scored. Name the group and the supervisor whose methods your project needs.
MEXT and government research scholarships
A field-of-study plan is required and is closer to a structured form than an essay. Follow the official template exactly and keep the language accessible to non-specialist selection panels.
Grant and fellowship applications
Budget justification is assessed alongside the science. Build costs against the funder's eligible categories from the start, because ineligible costs are removed rather than negotiated.
Connected to the live catalogue24 current relevant opportunities

This is a broader relevance count across the guide's linked opportunity collections, not a claim that every record explicitly requires this document or action.

Opportunities that need this

Official sources

Common questions

Frequently asked questions

How long should a research proposal be?

Follow the funder's limit. Doctoral applications typically ask for 1,500–3,000 words; fellowship and grant proposals run considerably longer. Exceeding a stated limit risks administrative rejection before assessment.

How specific does my proposal need to be if the topic might change?

Specific. Reviewers know projects evolve, and they are assessing whether you can design research, not holding you to the plan. A vague proposal reads as an inability to design, not as flexibility.

Should I contact a supervisor before writing the proposal?

Yes, in almost every system where doctoral funding is project-linked or nomination-based. A supervisor will tell you whether the question is feasible in their group before you spend weeks on it.

Do I need preliminary results?

Not for doctoral applications. For postdoctoral fellowships and grants, pilot data substantially strengthens a proposal because it evidences feasibility rather than asserting it.

Use the official instructions first.

Scholarship Signalguides support preparation. The provider's current application rules, templates and limits remain authoritative, and they change between cycles. How this guidance is written and reviewed