How to Write a Research Proposal: The Question, the Gap, and the Plan

A research proposal is not a literature review with a timeline attached.

It is a plan that asks a reader to believe three things: the question is worth answering, the method can answer it, and you can finish it with the time and resources you have. If any one of those is missing, the proposal is not ready, no matter how many references it contains.

This guide is for Indian students writing a research proposal for a master's dissertation, an MPhil, a PhD application or a funded project. Your department's format and your guide's expectations override this article. The structure below is a way to make the plan visible.

What a Research Proposal Is Actually For

A proposal has a different job from a paper.

A paper reports what happened. A proposal argues what should happen next. The reader is not asking "what did you find?" They are asking "is this a question worth funding, supervising and spending two years on?"

A proposal is also a genre with a reader in mind. In India, that reader may be a departmental committee, a guide, an ethics committee or a funding panel. Each one reads for a different risk: the committee asks whether the project fits the programme, the guide asks whether it can be supervised, the ethics committee asks whether participants are protected, and the funder asks whether the money buys a credible answer. Write for the most sceptical reader in the room.

That changes the writing. You do not need to prove your conclusion. You need to show that the question is answerable, the method is credible, and the plan is honest about its limits.

Purdue OWL's guide to academic proposals describes the proposal as a document that defines a problem, explains why it matters, and presents a plan for addressing it. Purdue's research proposal guide adds the practical test: the reader should be able to see what you will do, how you will do it, and why the project is feasible. That is a better checklist than "does it sound academic?"

The Seven-Part Structure

The title deserves its own pass. A good title names the question and the setting without turning into a sentence. "How First-Year Engineering Students Interpret Written Feedback: A Qualitative Study in Two Indian Departments" tells the reader what, who and how. "A Study of Feedback" tells them almost nothing. Write the title last, after the question is stable.

The seven parts are not seven equal sections. The question and method usually take the most space; the title and contribution can be short.

A worked example. Suppose you want to study how first-year engineering students use feedback. "Feedback in engineering education" is a field, not a question. "How do first-year engineering students interpret written feedback on their lab reports?" is a question. The gap might be that existing studies measure whether feedback was given, not how students read it. The method might be 12 interviews plus a review of the feedback sheets. The feasibility check asks whether you can access two departments and whether 12 interviews are enough for a qualitative study. Each part shapes the next.

A useful discipline: write the question in one sentence before you write the gap. Then test it against the method. If the method cannot produce evidence that answers the question, change one of them now, not after the fieldwork.

The Gap, the Method, and the Feasibility

The gap is a missing answer, not a missing topic

"No one has studied rural students in this district" is a topic gap. It does not tell the reader why the answer matters. A stronger gap is: "Existing studies measure access to coaching, but not whether the coaching changes the way students choose their subjects." Now the gap is a missing explanation, and the method can be designed around it.

Use three signals:

• the authors' own limitations section;
• a recommendation for future research;
• two studies that answer the same question differently.

If you cannot point to a sentence in a paper, the gap is still your hunch. That is allowed, but label it as a hunch and explain how you will test it.

The method must fit the question

A question about "how students experience feedback" is not answered by a survey that only asks "are you satisfied?" A question about "how many students received feedback" is not answered by six interviews.

Write a method table:

If a row has no data source, the method is not ready. If a row has a source but the source cannot be accessed, the method is also not ready. Access is part of the method, not a administrative detail you add later. If a row has data but no analysis, you have collection, not research.

Feasibility is part of the argument

Examiners reject good questions for bad logistics. Show:

• Access: do you have permission to collect this data?
• Time: can you recruit, collect and analyse within your programme?
• Skills: can you do the analysis, or will you need training?
• Cost: what does the fieldwork cost, and who pays?
• Risk: what happens if access is denied or the response rate is low?

A smaller, well-supported project is stronger than an ambitious one that cannot be finished. A useful test: write the first week of fieldwork, not the first year. Who do you contact on Monday? What document do you need? What happens if the answer is no? If you cannot write that paragraph, the plan is still abstract.

A smaller, well-supported project is stronger than an ambitious one that cannot be finished. If your plan depends on one office giving you data, say what you will do if they say no.

Ethics is not a paragraph you paste

Name the actual risks: consent, anonymity, power imbalance, sensitive questions, data storage. Say how you will reduce them. If your project involves minors, patients or vulnerable groups, follow your institutional ethics process before you collect anything.

Where Sovi.AI Fits, and What It Cannot Decide

Sovi Smart Writing can help you turn a proposal outline into readable prose. It cannot choose your question, find your gap, or promise that your method is feasible.

Smart Writing accepts a written requirement, attached files and a tone. At the time of writing, the requirement box takes up to 2,000 characters, the upload accepts PDF, Markdown and TXT, and the tone list includes Academic. You can add one to three previous documents as style samples. The draft appears in the editor, and you copy it into your own document.

Do not ask the tool to "find a research gap" from a broad field. It will produce a plausible gap, not a real one. Ask it to test whether the gap you found is stated clearly, or to point out where the method does not answer the question. That is a smaller, more useful job.

Use it after you have written the question, the gap and the method table. A prompt that keeps the boundary:

You are editing my research proposal, not choosing my research question.
Question: [one sentence].
Gap: [what is missing, with the source].
Method: [data, sample, analysis].
Feasibility: [access, time, skills, cost].
Rewrite the gap and method sections in 250 words.
Use only the sources I list. Mark every missing citation as [SOURCE NEEDED].
Do not invent a finding, a sample size or a permission.

The output is a draft. Check every citation, every claim about the literature and every promise about access. The tool does not know your department's ethics process or your supervisor's threshold for "feasible."

The research proposal kit

The kit below turns the seven parts into a one-page plan, a method table and a feasibility check and sources.

• research-proposal-kit.pdf to print and fill in by hand.
• research-proposal-kit.docx to type into and adapt to your programme.

Inside:

1. The one-page plan, with the question, gap, method and contribution.
2. The method table, with data, source, analysis and risk.
3. The feasibility and ethics check, with access, time, cost and consent.

Fill the one-page plan before you write the introduction. Then give it to someone outside your field and ask them to repeat the question and method back to you. If they cannot, the plan is not clear yet, even if the prose is polished.

Fill the one-page plan before you write the introduction. If the question and method do not fit on one page, the proposal is still a set of ideas.

Frequently asked questions

1. How long should a research proposal be?

Your programme decides. A short internal proposal can be 1,500–2,000 words; a PhD application may ask for 2,000–3,000 or a specific page limit. Check the call before you write. Do not pad the literature review to reach a page count.

2. Do I need a hypothesis?

Not always. A hypothesis is useful when you are testing a specific relationship. Exploratory or qualitative proposals may use research questions instead. Ask your guide which convention your department expects.

3. How many references should I include?

Enough to show that you know the conversation around the question, not a bibliography. A short proposal may use 15–25 sources; a longer one may use more. Every source should do work in the argument.

4. What if my proposed method is too expensive or too slow?

Change the scope, not the honesty. A smaller sample, a shorter timeframe or a secondary dataset can make the project feasible. Explain the trade-off and what it costs in terms of what you can claim.

5. Can I use AI to write the proposal?

Use it to improve the prose, tighten the gap or check whether the method section is clear. Do not use it to invent a citation, a finding, a sample or a permission. You will have to defend every claim in the proposal.

6. What is the most common reason a proposal is rejected?

A mismatch between the question, the method and the resources. The question is interesting, but the method cannot answer it, or the plan cannot be completed with the access and time available. Fix the mismatch before you polish the wording.