How to structure a research paper: the IMRaD format explained
On this page
- The logic behind IMRaD: the hourglass
- Introduction: from problem to question
- Literature review or background
- Methods: enough detail to repeat the study
- Results: what you found, without interpretation
- Discussion: what it means
- Conclusion
- The parts around the core
- How long should each section be?
- When IMRaD does not fit
- A practical way to start
- Quick checklist
Most empirical papers in the sciences, social sciences, medicine and engineering follow the same skeleton: Introduction, Methods, Results and Discussion, usually shortened to IMRaD. It is not a bureaucratic template. Each part answers one question a reader has, in the order they tend to ask it: Why does this matter? What did you do? What did you find? What does it mean? Once you understand the job of each section, the structure stops feeling like a form to fill in and starts working as a thinking tool.
This guide walks through each section, gives rough proportions, and covers the parts that sit around the core: title, abstract, literature review, conclusion and references.
The logic behind IMRaD: the hourglass
A useful way to picture a research paper is an hourglass. The introduction starts broad (a real-world or theoretical problem) and narrows to your specific question. The methods and results are the narrow neck: concrete, specific and limited to your study. The discussion widens again, moving from your findings back out to the field, theory and practice.
If a reader can follow that shape, they can read your paper quickly. If your introduction wanders between levels, or your discussion introduces new data, the hourglass breaks and the reader has to work harder.
Introduction: from problem to question
The introduction has three jobs, often described as "establish a territory, find a niche, occupy the niche":
- Establish the territory. Show that the topic matters and summarize what is already known. Keep this tight; it is context, not a full review.
- Identify the gap. Point to what is missing, contested or unexplained. A good gap is specific: "few studies have examined X among Y using Z" is better than "more research is needed".
- State your contribution. Give your research question, aims or hypotheses, and briefly say how you address them. Many fields also expect a one-paragraph roadmap of the paper.
A common mistake is writing a history of the field before reaching the point. If your research question does not appear until page four, move it up. Readers should know what the paper is about by the end of the first page.
Literature review or background
In many disciplines the literature review is a separate section between the introduction and the methods; in others it is folded into a longer introduction. Either way, its purpose is to build the case for your study, not to prove you have read everything. Organize it by themes, debates or variables rather than by author, and end it by showing how the gaps lead to your hypotheses. Our guide to writing a literature review covers this in detail.
Methods: enough detail to repeat the study
The methods section explains exactly what you did so that a competent reader could evaluate or replicate it. Typical subsections include:
- research design (experiment, survey, case study, secondary data analysis, and so on)
- participants or data sources, sampling and inclusion criteria
- materials, instruments and measures, with reliability or validity evidence
- procedure, in the order things happened
- analysis strategy, including software and the main statistical or qualitative techniques
- ethics approval and consent, where relevant
Write it in past tense for completed work and justify choices that a reviewer might question. See the methodology section guide for a section-by-section checklist.
Results: what you found, without interpretation
Results report findings in a logical order, usually following your research questions or hypotheses. Lead with the finding, then support it with the numbers or evidence: "Remote-work intensity was positively associated with exhaustion (β = …, p = …)". Use tables and figures for detail and let the text highlight what matters most. Avoid repeating every number from a table in prose.
Keep interpretation out of this section. "This suggests that managers should…" belongs in the discussion. If you are writing before your data are collected, draft the results section as a template: planned analyses, empty table shells and clearly marked placeholders such as [insert coefficient]. Never fill these with invented values.
Discussion: what it means
The discussion reverses the introduction's funnel. Start by restating the main answer to your research question in plain words, then compare your findings with prior work, explain unexpected results, and spell out theoretical and practical implications. Close with limitations and directions for future research. Our discussion section guide breaks this into a paragraph-by-paragraph plan.
Conclusion
Some journals merge the conclusion into the discussion; many theses and course papers ask for a separate one. A conclusion is short: the main takeaway, why it matters, and what should happen next. It should not introduce new evidence or citations you have not discussed.
The parts around the core
- Title. Specific and searchable. Include the main variables, population or context, and sometimes the design: "Remote Work Intensity and Burnout Among Knowledge Workers: A Two-Wave Survey".
- Abstract. A 150–300 word summary of the whole paper, usually written last. See how to write an abstract.
- Keywords. Four to six terms a researcher would actually search for.
- References. Every in-text citation must appear in the list and vice versa. See APA vs author-year citation basics.
- Appendices. Survey instruments, extra tables and robustness checks that would interrupt the main text.
How long should each section be?
Proportions vary by field and journal, but for a typical empirical paper of 8,000–10,000 words, a reasonable starting budget is:
- Introduction: 10–15%
- Literature review or background: 15–25%
- Methods: 15–20%
- Results: 20–25%
- Discussion and conclusion: 20–25%
Check two or three recent papers in your target journal or your department's thesis guidelines and adjust. If your methods section is longer than your discussion, you may be describing procedure in too much detail or under-interpreting your findings.
When IMRaD does not fit
IMRaD suits empirical studies. Other paper types use different structures:
- Review articles organize the body by themes, methods or chronology and often add a methods section describing the search strategy.
- Theoretical or conceptual papers build an argument through a sequence of propositions.
- Humanities papers usually follow an argument-driven structure with thematic sections rather than fixed headings.
- Case studies and design papers often use problem, context, intervention, outcome and lessons.
Even then, the underlying logic still helps: context, gap, approach, findings, meaning.
A practical way to start
Before writing paragraphs, write one sentence for each section: the problem, the gap, your question, your design, your expected or actual main finding, and its main implication. If those six sentences make sense together, you have a paper. Then expand each into a bullet outline with the sources you will cite, and only then start drafting. Our guide on turning notes into a first draft shows how to move from that outline to full prose.
Quick checklist
- Can a reader find your research question on the first page?
- Does the literature review lead logically to your hypotheses or aims?
- Could someone repeat your study from the methods section?
- Are the results free of interpretation and fabricated values?
- Does the discussion answer the question you asked in the introduction?
- Do the title and abstract match what the paper actually delivers?