How to write a research proposal — structure, length and example (2026)

Before anyone lets you write a thesis, you usually have to prove the thesis is worth writing. That proof is the proposal, and knowing how to write a research proposal is the difference between a topic approved in one meeting and a topic sent back three times. A proposal is a short, structured argument with four moving parts: here is a gap in what we know, here is the question that closes it, here is the method that answers the question, and here is the evidence I can finish it in the time I have. Supervisors read dozens of these each semester and reject most on feasibility, not on ambition. This guide covers the eight standard sections, realistic word counts by degree level, how to phrase questions that survive review, a timeline table you can copy, and the five failure patterns that get proposals returned.

What a research proposal is — and what it is not

A research proposal is a forward-looking document. Everything in it is in the future tense: what you will investigate, what data you will collect, what you expect to find. This single distinction resolves most of the confusion students bring to their first draft.

It is not a literature review, although it contains a compressed one. It is not an abstract, which summarises work already done. It is not the first chapter of your thesis, although perhaps 40% of the text will survive into your thesis introduction in revised form. The proposal has one job: convincing a reader with limited time that the project is worth supervising and that you are capable of delivering it.

That audience matters. Your reader is not a general academic — it is one specific person who will be responsible for your progress for the next eight to eighteen months. They are asking three questions while reading: is this original enough to defend, is it narrow enough to finish, and does this student understand what the method actually requires? Write to those three questions and the document almost structures itself.

Note: Many departments publish their own proposal template with mandatory headings and a word limit. If one exists, it overrides every general guideline in this article. Ask for it before you start writing — rewriting a finished proposal into someone else’s section headings costs a full evening.

Research proposal structure — the eight standard sections

Formats vary by discipline, but almost every proposal template is a variation on the same eight blocks. Below is the sequence, what belongs in each, and how much space to give it in a 2,000-word bachelor’s proposal.

Section	Purpose	Share of word count
Title	Names the variables, population and context	15–20 words
Background	Establishes the field and the state of knowledge	20%
Problem statement	Names the specific gap	10%
Aims and objectives	States what the study will achieve	8%
Research questions / hypotheses	The questions that will be answered	8%
Methodology	Design, sample, instruments, analysis	30%
Timeline	Month-by-month schedule to submission	7%
References	15–30 sources, consistently formatted	12%

Title, background and problem statement

A working title should contain the independent variable, the dependent variable, and the population or setting. “Remote work and employee engagement” is a subject, not a title. “The effect of hybrid work arrangements on employee engagement in Polish IT companies, 2023–2026” is a title — a reader knows immediately what will be measured, in whom, and when.

The background section compresses the literature review into three to five paragraphs. Do not narrate every paper you have read. Group sources into two or three positions in the field, show where they agree, and end on the point where they stop. That endpoint is your gap. The technique is the same one used for a full literature review chapter, just executed at one-tenth the length.

The problem statement is one paragraph, sometimes three sentences. It says what is not known, why that absence matters practically or theoretically, and what would change if it were known. Weak proposals describe a topic here. Strong proposals describe a deficit.

Aims, objectives and research questions

Aims and objectives are not synonyms, and supervisors notice when they are treated as such. The aim is singular and broad: “to determine whether hybrid work arrangements affect engagement.” Objectives are three to five concrete steps toward it: to measure engagement across two work models, to identify moderating variables, to compare results against the Utrecht Work Engagement Scale norms.

Research questions then convert objectives into answerable form. Each objective should map onto exactly one question or hypothesis. If you have four objectives and seven questions, something is unaccounted for — and a reviewer will find it.

Methodology, timeline and references

The methodology section is where proposals are won or lost, and it deserves the largest share of your word count. State the design (quantitative, qualitative, mixed), the sampling strategy and target size, the instrument, the data collection procedure, and the specific analysis you will run. “Statistical analysis will be performed” is not a method. “Independent-samples t-tests will compare mean engagement scores between the two groups, with Cohen’s d reported for effect size” is. A fuller treatment of these choices belongs in the methodology chapter, but the proposal must already name them.

References should be complete and consistently formatted from the first draft. Use whatever style your faculty mandates — the official APA Style guidance covers the seventh edition rules for most social science departments. Fifteen sources is thin for a bachelor’s proposal; thirty is comfortable. Half should be from the last five years.

How long should a research proposal be?

Length depends on degree level and on whether the proposal is an internal approval document or a funding application. The table below reflects typical European requirements.

Context	Word count	Pages (1.5 spacing)	References
Bachelor’s thesis proposal	1,500–2,500	4–7	15–25
Master’s thesis proposal	3,000–5,000	8–14	30–60
Doctoral / PhD proposal	6,000–10,000	15–25	60–120
Grant or funding application	Per call template	Per call template	Per call template

Funding applications are a separate genre with binding page limits and scoring rubrics — the Horizon Europe programme documentation shows how prescriptive those templates become. For a degree proposal, the honest rule is: write to the shortest length that fully specifies the method, then stop.

How to phrase research questions a supervisor will accept

Most rejected proposals fail here, and the failure is usually mechanical rather than conceptual. A question is acceptable when it is answerable with data you can realistically obtain, when the variables in it are measurable, and when it is narrow enough that one study can close it.

The four tests every question must pass

Run each draft question through these four checks before the proposal leaves your desk.

Measurability — can every noun in the question be operationalised? “Does organisational culture improve wellbeing?” fails; neither term names an instrument. “Does perceived organisational support, measured by the SPOS scale, predict scores on the WHO-5 Wellbeing Index?” passes.
Scope — could the question be answered by one study of the size you are proposing? If answering it honestly requires a five-year longitudinal design, it is a research programme, not a thesis.
Access — do you have a realistic route to the respondents, documents or dataset? A question about board-level decision-making in listed companies fails on access for most undergraduates.
Non-triviality — is the answer already known? If a 2024 meta-analysis settled it, you are replicating, not researching, and you must say so explicitly.

Rewriting weak questions into strong ones

Weak question	Why it fails	Rewritten
Is social media bad for students?	Both terms unmeasurable, value-laden	Does daily Instagram use above 120 minutes correlate with lower GPA among second-year students at one faculty?
How can companies improve retention?	Prescriptive, unbounded	Which of four retention practices do IT employees in Kraków rate as most influential on intention to stay?
What is the role of AI in education?	No variable, no population	How do secondary school teachers in Lesser Poland assess the accuracy of AI-generated feedback on student essays?

Feasibility and the timeline table

Feasibility is the section students skip and reviewers read first. A timeline demonstrates that you have converted an idea into a sequence of tasks with dates. Present it as a table rather than prose — it reads faster and it is much harder to fake.

Month	Milestone	Deliverable
1	Literature search and screening	25 sources catalogued in Zotero
2	Theoretical chapter draft	12–15 pages
3	Instrument design, ethics approval	Questionnaire + consent form
4	Data collection	Minimum 120 responses
5	Analysis	Coded dataset, test results
6	Empirical chapter and discussion	20 pages
7	Full draft to supervisor	Complete manuscript
8	Revisions and submission	Final version, antiplagiarism check

Build the table backwards from your submission deadline, not forwards from today. Then add a 20% buffer to data collection specifically — it is the phase that slips, because it depends on other people responding. Note any dependency that is outside your control: ethics committee sittings, institutional access permissions, seasonal availability of respondents.

Drafting a proposal to this level of specificity takes most students two to three working days, and much of that time goes into structural scaffolding rather than thinking. If you would rather spend those days on the substance, the Smart-Edu bachelor’s thesis generator produces a complete structured draft with a bibliography in 30–90 minutes from 249 PLN, following the same section logic described above — leaving you to verify the sources and sharpen the argument.

Why proposals get rejected

Five patterns account for the large majority of returned proposals. None of them are about intelligence.

Scope too broad. The single most common reason. “The impact of digitalisation on the Polish economy” cannot be answered by one person in six months.
No identifiable gap. The background describes a field but never states what is missing, so the reader cannot see why the study should exist.
Method mismatched to question. A qualitative question with a survey instrument attached, or a causal claim proposed on cross-sectional correlational data.
Unrealistic sample. A promised 400 respondents with no described route to reaching them, or interviews with executives who have no reason to agree.
No ethics plan where one is required. Studies involving minors, patients or sensitive personal data need committee approval, and a proposal that ignores this will be returned regardless of quality.

Tip: Before submitting, hand the proposal to a peer outside your specialism and ask them to state your research question in their own words. If they cannot, the question is not yet clear enough for a reviewer either.

Frequently asked questions about research proposals

How long does it take to write a research proposal?

For a bachelor’s proposal, budget 15–20 hours spread over two weeks: roughly six hours for the literature search, four for drafting the background and problem statement, five for the methodology, and the rest for the timeline, references and revision. The literature search is what expands unpredictably — tools like Google Scholar shorten it, but screening abstracts still takes real time.

Do I need results or preliminary data in a proposal?

No, and including invented ones is a serious error. A proposal is written before data collection. What you may include is a pilot result if you genuinely ran a pilot, clearly labelled as such, or descriptive statistics from an existing public dataset you intend to analyse.

Can I change my topic after the proposal is approved?

Minor adjustments to scope, sample size or instrument are normal and expected — most theses drift somewhat from their proposal. Changing the research question entirely usually requires a new approval, and at some institutions a formal re-submission. Raise it with your supervisor as soon as you suspect the original question will not work, not three months later.

What if my supervisor rejects the proposal?

Ask for the specific reason and treat it as diagnostic. Rejections for scope are the easiest to fix — narrow the population or the time frame and resubmit within a week. Rejections for method mismatch require rethinking the design. Rejections for lack of a gap mean returning to the literature. In all three cases the second version is typically approved, so the cost is time rather than the topic itself.

Should the proposal use the same citation style as the thesis?

Yes. Use the style your faculty requires from the first draft, and keep the reference manager library you build during the proposal — it becomes the foundation of your final bibliography. Reformatting 60 references at the end of the thesis is an avoidable evening.

Summary

Knowing how to write a research proposal comes down to controlling four things: a gap stated explicitly rather than implied, a question narrow enough to answer with the resources you have, a method described in enough detail that a stranger could execute it, and a timeline that survives contact with a calendar. Get those right and the eight sections largely write themselves. The proposal that gets approved fastest is rarely the most ambitious one — it is the one where a supervisor can see, in ten minutes of reading, exactly what will be done and exactly when it will be finished. Start by choosing a topic that passes the four question tests, and the rest of the document becomes an exercise in specification rather than persuasion.