Written and maintained by CASRAI Editorial Board
Published · Last updated
On this page
An IMRaD structure worked example applies the four sections to one paper: the Introduction says why the study was done, Methods says what was done, Results says what was found, and the Discussion says what it means. This guide builds that outline for a single illustrative study, then adds a sentence-sorting exercise, a blank template, the ICMJE rules for each section, and a comparison of how journals place Methods. The definitions are in CASRAI’s IMRaD Structure dictionary entry.
Illustrative example, not a real study. The paper outlined below — a fictional study of a brief mindfulness intervention and graduate-student self-reported stress — is a composite constructed to demonstrate structure only. No institution, dataset, participant count, statistic, or finding described here is real or attributed to any actual publication. Do not cite it. Its only purpose is to show, section by section, how IMRaD organizes an empirical paper’s content.
Why a worked example, not just definitions
Knowing that “Methods describes what you did” and “Results reports what you found” is necessary but not sufficient — the failure mode most first-time authors hit isn’t misunderstanding the definitions, it’s not seeing how a specific claim, sentence, or piece of data actually gets assigned to one section and not another when it’s genuinely ambiguous which section it “belongs” to. A worked example makes that assignment visible. The outline below is deliberately modest in scope — a single-study, single-method paper — because the goal is to show the mechanics of the structure clearly, not to model every possible variation covered in CASRAI’s IMRaD Structure entry (structured abstracts, multi-study Results/Discussion pairs, field-specific naming like “Materials and Methods”).
What the ICMJE Recommendations say each section does
The International Committee of Medical Journal Editors (ICMJE) describes the IMRAD structure as “not an arbitrary publication format but a reflection of the process of scientific discovery” in its guidance on preparing a manuscript for submission. Its rules map directly onto the outline above, and they are the quickest way to settle a borderline placement decision.
- Introduction. Provide the context or background for the study, then state the specific purpose, research objective or hypothesis tested. Cite only directly pertinent references, and do not include data or conclusions from the work being reported.
- Methods. Make clear how and why the study was done, with enough detail that others could reproduce it. Include only information that was available at the time the plan or protocol for the study was written.
- Results. All information obtained during the study belongs here. Present main findings first, report all prespecified primary and secondary outcomes, and do not repeat in the text every number that appears in a table. Give absolute numbers alongside percentages.
- Discussion. Summarise the main findings, discuss possible mechanisms or explanations, emphasise what is new, and state the limitations of the study. Discuss implications, link conclusions to the study goals, and avoid repeating material from the Introduction or Results. Separate clinical significance from statistical significance.
ICMJE also asks that statistical reporting go beyond hypothesis tests alone: use confidence intervals where possible, and distinguish prespecified analyses from exploratory or subgroup analyses. In the illustrative paper, this means the Results would state the prespecified paired comparison first and label any subgroup pattern as exploratory.
The illustrative paper: setup
Working title: “A Brief Mindfulness Intervention and Self-Reported Stress Among First-Year Graduate Students: A Pilot Study” (illustrative only).
Premise: A researcher wants to know whether a short, structured mindfulness exercise, done daily for two weeks, is associated with a change in first-year graduate students’ self-reported stress, measured with a standard validated stress scale before and after the intervention period. To plan a study like this before writing it up, see how to write a paper proposal.
Section-by-section outline with representative content
Title and abstract
The title names the intervention, the outcome, and the population — no results are previewed. The abstract (structured or unstructured, per the journal’s house style — see CASRAI’s structured abstract format guide) restates purpose, method, principal result, and conclusion in miniature, positioned before the Introduction and outside the four core IMRaD sections proper. For the page that precedes the paper in APA style, see the APA cover page (title page) guide.
1. Introduction
Representative content, in order:
- Broad context: One or two sentences establishing that graduate-student mental health and stress are an active area of concern in higher education, citing prior survey or prevalence literature (real citations would go here — omitted in this illustrative outline).
- Narrowing to the gap: A sentence identifying what prior work has and hasn’t established — e.g., that brief, low-cost mindfulness interventions have shown promise in other populations, but evidence specific to first-year graduate students transitioning into a new program is limited.
- The specific aim: A single explicit sentence stating the study’s objective or hypothesis — e.g., “This pilot study tested whether a two-week, ten-minute daily guided mindfulness exercise is associated with a reduction in self-reported stress among first-year graduate students, measured using [a named, validated stress scale].”
What does NOT belong here: any statement of what the study found. “We found that stress scores decreased” is a Results/Discussion sentence — in the Introduction, the closest permissible framing is only the stated hypothesis or expectation, not the outcome.
2. Methods
Representative content, in order:
- Participants: Who was studied, how they were recruited, and eligibility criteria — e.g., “Participants were first-year students (n = [illustrative figure]) enrolled in [illustrative department], recruited via a program-wide email in the first week of the term.”
- Materials/instrument: The validated stress scale used, with a citation to the instrument’s source and any reported reliability statistic from that source — not from this study’s own data (that would be a Results item).
- Procedure: Exactly what participants did and when — baseline survey administration, the daily mindfulness exercise (format, duration, delivery mechanism), the follow-up survey timing.
- Analysis plan: The statistical test used to compare pre- and post-intervention scores, stated as a plan, not a result — e.g., “Paired scores were compared using [named test].”
What does NOT belong here: the outcome of running that test. A common misplacement is writing “…compared using [named test], which showed a significant decrease” — the test result belongs in Results; Methods states only that the test was planned and run.
3. Results
Representative content, in order:
- Sample description: Who actually completed the study — e.g., final n, any attrition between baseline and follow-up, basic descriptive statistics of the sample.
- Primary finding, stated neutrally: “Mean self-reported stress scores were [illustrative baseline value] at baseline and [illustrative follow-up value] at follow-up, a difference of [illustrative effect size] ([test statistic], p = [illustrative value]).” No adjective characterizing the finding as meaningful, surprising, or consistent with prior work — that framing belongs in the Discussion.
- Secondary/exploratory observations, if any: stated with the same neutrality — e.g., a subgroup pattern noted but not yet interpreted.
- Typically organized around the paper’s tables and figures, referenced in text (e.g., “as shown in Table 1”).
What does NOT belong here: interpretation. “This suggests the intervention was effective” or “this is consistent with [prior study]” are Discussion-section sentences; the single most common IMRaD misplacement error, per CASRAI’s IMRaD Structure entry, is exactly this — interpretive language leaking into Results.
4. Discussion
Representative content, in order:
- Restate the main finding in plain language and connect it directly back to the aim stated in the Introduction — “Consistent with the study’s hypothesis, self-reported stress scores decreased over the two-week intervention period.”
- Interpretation: what the finding plausibly means, and how it relates to prior literature cited in the Introduction — where it agrees, extends, or diverges from what’s already known.
- Limitations: stated specifically, not generically — e.g., small pilot sample, no control group, self-report-only outcome measure, short follow-up window. A generic “further research is needed” without naming the actual limitation is a common weakness reviewers flag.
- Implications and future work: what a larger, controlled follow-up study would need to establish, and what (if anything) the pilot result suggests for practice in the interim.
- Conclusion: a short closing statement — sometimes a separate final heading, sometimes the Discussion’s last paragraph, depending on journal convention (see the naming-variant discussion in CASRAI’s IMRaD Structure entry).
What does NOT belong here: new data. Every number discussed in the Discussion should already have appeared in Results — the Discussion interprets, it does not introduce.
The same outline, condensed as a checklist
| Section | Answers the question | Tense | Contains new data? | Contains interpretation? |
|---|---|---|---|---|
| Introduction | Why did you do this study? | Present/past mixed | No | No (only the stated hypothesis) |
| Methods | What did you do? | Past | No | No |
| Results | What did you find? | Past | Yes | No |
| Discussion | What does it mean? | Present/past mixed | No (only reference to data already reported) | Yes |
Where the Methods and Results boundary falls
The ICMJE rule gives a testable criterion: if the information was available when the protocol was written (the eligibility criteria, the planned instrument, the planned analysis), it is a Methods item; if it was obtained during the study (how many people actually enrolled, how many dropped out, the observed scores), it is a Results item. In the illustrative paper, “participants were first-year graduate students recruited by email” is Methods, while “of the students who enrolled, some did not complete the follow-up survey” is Results, because attrition is only known after the study runs. The same logic explains why a scale’s published reliability belongs in Methods, with a citation, while the reliability observed in the study’s own data belongs in Results.
Sentence-sorting exercise on the illustrative paper
Placement errors are easiest to spot sentence by sentence. Each sentence below is invented for the illustrative paper; the table gives the section it belongs in and the rule that decides it.
| Sentence | Section | Deciding rule |
|---|---|---|
| Graduate-student stress is a recognised concern in higher education. | Introduction | Context and background for the study |
| We hypothesised that a daily ten-minute exercise would be associated with lower stress scores. | Introduction | Purpose or hypothesis tested, with no data |
| Students were eligible if they were in their first term of a graduate programme. | Methods | Known when the protocol was written |
| Paired scores were compared with a paired t-test. | Methods | A planned analysis, stated without its outcome |
| Fewer students completed the follow-up survey than the baseline survey. | Results | Information obtained during the study |
| Mean scores fell between baseline and follow-up (values and test statistic in Table 2). | Results | Main finding stated neutrally, pointing to the table |
| A pattern in one subgroup was noted but was not prespecified. | Results | Exploratory analyses are labelled as such |
| This is consistent with the hypothesis and may reflect reduced rumination. | Discussion | Interpretation and possible mechanism |
| Without a control group, the change cannot be attributed to the exercise. | Discussion | Limitations are stated in the Discussion |
| A larger controlled trial should test whether the change persists. | Discussion | Implications and future work |
Blank IMRaD template
This skeleton can be copied and filled in. The prompts are questions rather than prose so that the finished paper contains only the author’s own content.
- Title and abstract: Population, intervention or exposure, outcome. Abstract in the journal’s format.
- Introduction: What is known (context)? What is not known (gap)? What exactly did this study set out to test (aim or hypothesis)?
- Methods: Who or what was studied and how were they selected? What was measured, with which instrument or method? What was done, in what order? How were the data analysed (planned analyses only)? What approvals or registrations apply?
- Results: What was the final sample and what was lost along the way? What was the main finding? What were the other prespecified outcomes? What exploratory findings, labelled as exploratory, were there?
- Discussion: What was the main finding, in one plain sentence? How does it compare with the literature cited in the Introduction? What are the specific limitations? What follows for practice or future work?
- References and supplements: Only sources actually cited; supplementary material for detail that would interrupt the text.
Matching reporting guidelines to the sections
Reporting guidelines are checklists that tell authors which items to include and, in practice, which section carries them. The EQUATOR Network describes STROBE as reporting guidance for cohort, case-control and cross-sectional studies, and notes that it is not a tool for appraising study quality. STROBE was published in 2007 (Annals of Internal Medicine 2007;147(8):573-577) and in several other journals. Trials use the CONSORT checklist instead. Using the checklist for the study design while drafting helps decide, for each required item, whether it is a Methods item (design, eligibility, outcomes, analysis plan) or a Results item (participant flow, observed outcomes). The illustrative pilot has no control group, so it would be reported as a single-group before-and-after study and its limitations would say so.
How journals place Methods and combine sections
IMRaD is a default, not a universal template. The data table below the article compares three authoritative sources. In brief: the ICMJE recommendations describe Introduction, Methods, Results and Discussion in that order. PLOS ONE says the sections of a research article may be renamed and presented in any order, and that Results, Discussion and Conclusions may be combined. eLife’s standard research article puts Materials and Methods after the Discussion, while its medical submissions use the conventional order. An author should therefore draft in IMRaD logic and then arrange the sections as the target journal’s instructions require. The abstract limits in the table also differ, so check them before writing the abstract.
Common cross-section drafting mistakes, illustrated
- Result previewed in the Introduction: “Prior mindfulness interventions have shown mixed results, but as this study demonstrates, a brief intervention can meaningfully reduce graduate-student stress.” The clause after “but” belongs in the Discussion, not the Introduction — the Introduction should end at the stated hypothesis, not the outcome.
- Interpretation embedded in Methods: “Stress was measured using [scale], which is a highly sensitive instrument well suited to detecting the kind of change this study ultimately observed.” The second half assumes and states a result before Results exists — Methods should describe the instrument and procedure only.
- Interpretation embedded in Results: “Stress scores decreased significantly (p = [value]), confirming that brief mindfulness practice is an effective stress-reduction tool for graduate students.” Everything after the comma is Discussion content — Results should stop at the neutral statistical statement.
- New data introduced in the Discussion: “As discussed above, and as a follow-up analysis not shown here further confirms, the effect appears robust.” Any data supporting a claim needs to have already appeared in Results (with its own table/figure), not be introduced for the first time in the Discussion.
How this differs across paper types
This worked example is a single-study empirical pilot — the simplest case. Multi-experiment papers commonly pair a Methods/Results block per study under separate subheadings, then close with one integrated general Discussion; APA’s Publication Manual explicitly permits either that structure or a combined “Results and Discussion” heading per study, provided a comprehensive Discussion still closes the paper. Systematic reviews and meta-analyses use a related but distinct structure governed by PRISMA reporting guidelines, where “Methods” describes a search-and-screening protocol rather than an experimental procedure. See CASRAI’s IMRaD Structure entry for the fuller list of variants and where IMRaD does not apply (narrative reviews, case reports, most humanities argument essays).
Frequently asked questions
Does every empirical paper need exactly four sections in this order?
The four-part logic (why → how → what was found → what it means) is close to universal for original-research reporting in the sciences and social sciences, but exact section labels and counts vary: “Materials and Methods” is common in the life sciences, APA uses the singular “Method,” and some journals add a separate “Conclusion” heading after the Discussion. Always check the target journal’s specific author guidelines — IMRaD describes the dominant convention, not a universal mandate enforced identically everywhere.
Where does the Literature Review go in an IMRaD paper?
There typically isn’t a separate “Literature Review” section in a standard IMRaD empirical paper — the relevant prior literature is woven into the Introduction (to establish the gap the study addresses) and revisited in the Discussion (to interpret the findings against it). A standalone literature-review section is more characteristic of a narrative review article, which is not itself IMRaD-structured.
Can Results and Discussion be combined into one section?
Some journals and paper types allow or require a combined “Results and Discussion” section, particularly for multi-experiment papers reporting one study at a time. Where this is permitted, the discipline described above still applies within the combined section — each individual result should still be stated neutrally before being interpreted, even if both appear in immediate succession rather than under separate headings.
What goes in the Abstract if it summarizes all four sections?
A structured or unstructured abstract briefly restates the purpose (from the Introduction), the method (from Methods), the principal finding (from Results), and the main conclusion (from the Discussion) — see CASRAI’s structured abstract format guide for how this is organized under ICMJE, CONSORT, and PRISMA conventions specifically.
What does IMRaD stand for, and why is it used?
IMRaD stands for Introduction, Methods, Results and Discussion. The ICMJE describes it as a reflection of the process of scientific discovery rather than an arbitrary publication format, which is why readers can find the question, the procedure, the findings and the interpretation in predictable places.
Why do some journals put Methods after the Discussion?
Journal instructions set the order. eLife’s standard research article places Materials and Methods after the Discussion, whereas its medical submissions use Introduction, Methods, Results, Discussion. PLOS ONE allows sections to be presented in any order.
Where do limitations go?
In the Discussion. The ICMJE asks authors to state the limitations of the study there, and to name them specifically rather than as a generic call for further research.
Is information gathered during the study a Methods item?
No. Under the ICMJE rule, Methods holds only what was available when the protocol was written, and everything obtained during the study belongs in Results.
How long should the abstract be?
It depends on the journal. Per the data table below, PLOS ONE sets a maximum of 300 words, while eLife describes 150 to 200 words as normal and about 250 for medical submissions. The ICMJE text does not state a word limit.
Related CASRAI resources
- Dictionary: IMRaD Structure — the operational definition, origin (ANSI Z39.16, ICMJE), and full list of naming/structural variants.
- How to write a research paper
- Research paper format
- Writing the Results section of a research paper
- How to write the Discussion section of a research paper
- Structured abstract format
- Citation styles and referencing formats
Sources (accessed 2026-11-13)
- ICMJE Recommendations, Preparing for Submission: https://www.icmje.org/recommendations/browse/manuscript-preparation/preparing-for-submission.html
- PLOS ONE Submission Guidelines, Manuscript Organization: https://journals.plos.org/plosone/s/submission-guidelines
- eLife author instructions for research articles: https://elife-rp.msubmit.net/html/elife-rp_author_instructions.html
- EQUATOR Network, STROBE reporting guideline: https://www.equator-network.org/reporting-guidelines/strobe/
Further reading: IMRaD Structure — A section-by-section comparison of how humanities articles are organized (argument-driven, thematic or chronological) versus the IMRaD Introduction-Methods-Results-Discussion structure used in STEM, and why the two diverge.
Data: How three authoritative author guidelines position the IMRaD sections, Methods and abstract limits
Compares the section order, Methods placement, combined-section allowance and abstract and title limits stated in the ICMJE Recommendations, PLOS ONE submission guidelines and eLife author instructions. Compiled by reading each body's published guidance; cells say 'Not stated' where the source gives no figure.
Period covered: 2026.
| Source and article type | Default main-text section order | Methods placement | Results and Discussion may be combined | Abstract maximum (words) | Title limit (characters) |
|---|---|---|---|---|---|
| ICMJE Recommendations (IMRAD for original research) | Introduction, Methods, Results, Discussion | After Introduction | Not addressed | Not stated | Not stated |
| PLOS ONE research article | Introduction, then Materials and Methods, Results, Discussion, optional Conclusions (may be renamed and presented in any order) | Flexible (any order) | Yes | 300 | 250 |
| eLife research article (standard) | Introduction, Results, Discussion, Materials and Methods | After Discussion | Not stated | Not stated (150-200 words described as normal) | Up to 120 (suggested) |
| eLife research article (medical submissions) | Introduction, Methods, Results, Discussion | After Introduction | Not stated | About 250 | Not stated |
Sources
- ICMJE Recommendations: Preparing for Submission (IMRAD) (accessed 2026-11-13)
- PLOS ONE Submission Guidelines (Manuscript Organization) (accessed 2026-11-13)
- eLife author instructions (Research Articles) (accessed 2026-11-13)
Compiled by CASRAI from the sources above. Free to reuse under CC BY 4.0 with attribution.







