A practical SciCampus guide for researchers, PhD candidates, and academic authors
You have the data. You have the plots, the notes, the late-night doubts, and a growing suspicion that the hardest part of the project has only just begun: turning all of it into a paper that another scientist can understand and trust.
Here is the reassuring truth: you do not need to make your manuscript sound complicated to make it sound scholarly. You need to make its logic impossible to miss.
That is what IMRAD gives you. Introduction, Methods, Results, and Discussion are not four boxes you fill at the end of a study. They are a reader's path through your reasoning: Why did this question matter? What did you do? What did you find? What does it mean?
When that path is clear, reviewers spend their energy engaging with your science. When it is unclear, they start hunting for gaps, inconsistencies, and unspoken assumptions. This guide will help you build the first kind of paper.
Why Does IMRAD Still Work?
It Is a Trust-Building System
IMRAD has survived because it mirrors how scientific claims should be examined. Readers need context before they can judge relevance. They need methods before they can judge results. They need results before they can judge interpretation.
In other words: IMRAD separates what you asked, what you did, what you observed, and what you think it means.
That separation matters. Your Introduction cannot use your own unpublished findings as proof. Your Results should not quietly turn into an argument for policy. Your Discussion should not introduce a new analysis that was never explained in Methods. Each section has a job, and doing that job well makes the whole paper more credible.
The Four Sections Are Connected
Think of IMRAD as a chain, not a filing cabinet. The final paragraph of the Introduction promises a question. The Methods show how you answered it. The Results provide the answer. The Discussion explains its importance without stretching beyond the evidence.
A simple editorial test: if a reviewer cannot trace your conclusion back through a reported result, a described analysis, and a stated objective, the manuscript has a structural problem, not merely a stylistic one.
Use the Right Reporting Guide
IMRAD provides the backbone. Design-specific reporting guidelines provide the essential detail. For example, CONSORT supports reporting of randomised trials, STROBE supports observational studies, PRISMA supports systematic reviews, STARD supports diagnostic-accuracy studies, and TRIPOD supports prediction-model studies. The EQUATOR Network brings these and many other resources together.
Do not treat a checklist as paperwork you complete the night before submission. Use it early, while you are planning and drafting. It helps you notice the details a careful reader will need to evaluate the work: participant flow, missing data, protocol registration, model validation, or recruitment context.
Before You Draft: Find the Spine
A common writing mistake is to begin with the Introduction because it appears first. For many authors, that produces a polished background and a fuzzy paper. Start instead with the study's scientific spine.
What problem or uncertainty does this study address?
What specific gap remains after the most relevant evidence is considered?
What is your primary objective or research question?
What design, population, data, and outcome can answer that question?
What conclusion could this design support, and what conclusion could it not support?
Write the answers in plain language on one page. Keep that page beside you as you draft. It will prevent a very common problem: a manuscript that seems to be about one question in the Introduction, another in the Results, and a third in the Discussion.
Introduction: Earn the Reader's Attention
Your Introduction has one central purpose: convince the reader that this exact study needed to happen. It is not a history lesson. It is not a catalogue of everything ever published on the topic. It is a focused argument that leads to a clear, answerable question.
The Four Moves
Start with the real problem. Give readers enough context to understand why the issue matters.
Summarise the decisive evidence. Select the studies that shape the question; do not simply list studies.
Name the gap precisely. Explain what earlier research did not establish.
End with your objective, research question, and hypothesis where appropriate.
Show the Gap; Do Not Announce It
❌ Bad: “Many studies have investigated sleep and academic performance. However, more research is needed.”
✅ Better: “Prior cross-sectional studies link shorter sleep with lower academic performance, but they cannot establish whether within-student changes in sleep precede changes in examination performance. We therefore examined this association prospectively across one academic term.”
The stronger version tells the reader what is known, what is limited, and exactly how the present study responds. That is the difference between a broad topic and a publishable rationale.
Avoid the Literature-Review Trap
A literature review becomes a trap when it grows wider but not sharper. Five paragraphs that each begin with “Smith et al. found...” may demonstrate reading, but they do not demonstrate reasoning.
❌ Bad: “Study A found X. Study B found Y. Study C found Z.”
✅ Better: “Existing findings are mixed: studies with objective exposure measurement tend to report a stronger association than those relying on retrospective self-report. This pattern suggests that measurement error may explain part of the inconsistency.”
Editor's expectation: cite the evidence that moves your argument forward. A short, well-synthesised Introduction is often more persuasive than an exhaustive one.
Finish With an Objective Readers Can Test
Your final Introduction paragraph is a contract with the reader. Make it concrete enough that the Methods and Results can visibly fulfil it.
❌ Bad: “This study explores factors associated with patient outcomes.”
✅ Better: “We assessed whether baseline frailty was associated with 30-day hospital readmission among adults discharged after heart-failure admission, adjusting for prespecified demographic and clinical covariates.”
The second objective identifies the exposure, outcome, population, timeframe, and analytical intent. It gives the paper a centre of gravity.
Methods: Let Readers Audit Your Work
Methods are where trust is won or lost. A reviewer does not need to admire your methods; they need to understand them well enough to judge whether they can support your claim.
Write Methods for a sceptical but fair reader. Assume they will ask, “Could a different choice here have changed the result?” If the answer is yes, explain that choice clearly.
What a Complete Methods Section Covers
Design and setting: name the design, sites or data sources, and relevant dates.
Participants or units of analysis: explain eligibility, recruitment or selection, exclusions, and the final analytic sample.
Variables and measurement: define exposures, interventions, comparators, outcomes, covariates, instruments, timing, and units.
Procedures and materials: provide sufficient operational detail for evaluation and, where relevant, replication. For computational studies, report software, versions, parameters, code, and workflow.
Ethics and transparency: state approvals or exemptions, consent where relevant, registration or protocol information, funding, conflicts, and data/code availability as required.
Analysis: identify primary and secondary analyses, effect measures, models, assumptions, missing-data approach, multiplicity decisions, sensitivity analyses, and software versions.
Do Not Hide Important Decisions
Use citations for established methods and supplements for lengthy technical detail. But keep decisions that could change the interpretation in the main paper. If you changed the primary outcome, excluded a subgroup, selected a threshold, or chose an imputation approach, readers should not have to excavate a supplement to discover it.
❌ Bad: “Data were analysed using standard statistical methods.”
✅ Better: “We estimated adjusted risk ratios with modified Poisson regression, used multiple imputation for covariates with missingness, and repeated analyses in complete cases as a prespecified sensitivity analysis.”
The better sentence does not require pages of code. It gives the reviewer the decisions needed to assess the analysis.
Protect the Line Between Plan and Finding
Write what you did, not what you wish you had planned. If an analysis was exploratory, label it exploratory. If the protocol changed, say so and explain why. Readers are usually more comfortable with a transparent deviation than with a suspiciously perfect narrative.
Results: Put the Evidence First
Results should feel calm. This is not the place to sell the finding. It is the place to let the evidence speak in a sequence that directly answers the objective you promised.
A Reader-Friendly Sequence
Account for the sample: who or what entered the analysis, and what happened to exclusions or loss to follow-up?
Describe the starting characteristics needed to understand the sample, dataset, or experimental conditions.
Report the primary result first, using effect estimates, uncertainty intervals, denominators, and absolute numbers where useful.
Report secondary, sensitivity, subgroup, and exploratory analyses in clearly labelled order.
Report null, adverse, and inconvenient results as faithfully as favourable ones.
Let prose tell readers where to look. Let figures and tables, when used by the journal, carry dense numerical detail. Do not make readers decode the same result three times in a paragraph, a table, and a figure.
Use Language That Matches the Evidence
❌ Bad: “The intervention dramatically improved survival.”
✅ Better: “Thirty-day survival was 92% in the intervention group and 88% in the comparator group (risk difference 4 percentage points; 95% confidence interval, 1 to 7).”
The second version gives readers the finding, its size, and its uncertainty. It does not decide for them that the effect is “dramatic.”
A p-value is not an interpretation. Report the magnitude of the effect and its uncertainty. Then discuss practical, clinical, or theoretical importance in the Discussion.
Be especially careful with causal language. An observational association is not automatically an effect. Say “was associated with” unless your design and analysis justify stronger causal wording.
Discussion: Explain Without Overreaching
The Discussion is where you become your reader's guide. You show what the result adds, where it fits, and where it stops. Strong discussions sound confident because they are precise, not because they make the largest possible claim.
A Dependable Structure
Open with the main finding and answer the research question directly.
Place the finding alongside the most relevant evidence. Explain agreement, disagreement, and meaningful differences in design or context.
Offer plausible mechanisms where evidence supports them; label informed speculation as speculation.
Discuss strengths and limitations candidly, including the likely direction or consequence of important biases when possible.
State implications in proportion to the evidence — for research, practice, policy, theory, or method.
End with a concise conclusion that does not exceed the Results.
Limitations Are Not a Ritual Paragraph
Avoid writing “This study has several limitations” and then naming generic concerns without explaining their consequence. Tell the reader how a limitation may affect interpretation.
❌ Bad: “Our study was limited by its small sample size.”
✅ Better: “The modest sample size produced wide confidence intervals, so the data are compatible with both a small benefit and no meaningful difference. Larger, multicentre studies should estimate the effect more precisely.”
This approach does not weaken your paper. It shows that you understand its evidentiary boundaries.
Replace Advocacy With Calibrated Inference
❌ Bad: “These findings prove that the programme should be implemented nationally.”
✅ Better: “The findings support further evaluation of the programme in a larger implementation study; the single-site design and short follow-up limit conclusions about wider effectiveness.”
The goal is not to sound cautious for its own sake. The goal is to make claims at the level your design, analysis, and uncertainty can genuinely support.
What Mistakes Make Editors Pause?
The Paper Does Not Know What Question It Asks
A vague objective invites a diffuse Methods section and a selective Discussion. Choose one primary question and make every part of the manuscript serve it. You may report secondary and exploratory work, but label it clearly.
The Claim Is Stronger Than the Design
A cross-sectional study cannot quietly become a causal study in the Discussion. A narrowly selected dataset cannot become evidence about every population. Do not try to write around these limits. State them, then make the strongest accurate claim available.
The Numbers Do Not Travel Consistently
Editors notice when sample sizes, definitions, timeframes, or effect directions shift between the abstract, main text, figures, and supplement. Before submission, trace your primary outcome through every location where it appears.
Selective Reporting Damages Trust
If the Methods promise an outcome, Results should report it. If the analysis changed, disclose the change. If a result is null, it still belongs in the scientific record. Selective reporting may be more damaging than a modest or negative finding.
Formatting Comes Before Fit
A technically good paper can still receive a rapid rejection if it ignores the journal's scope, article type, word limit, data policy, or reporting-checklist requirement. Choose the target journal early and draft toward its instructions — our guide to choosing a Q1/Q2 journal by SJR and quartile covers how to narrow that choice before you draft.
Your Practical Submission Workflow
Choose the journal before the final draft. Read the author instructions, scope, article type, and reporting requirements.
Write a one-sentence primary objective and a one-sentence conclusion that the design could legitimately support.
Select the appropriate EQUATOR reporting guideline and map its items to the manuscript, supplement, or repository.
Draft Methods and Results together. Every planned analysis should have a Results placeholder; every reported outcome should have a Methods definition.
Write the Introduction once the scientific spine is stable. Write the Discussion once the results are fixed.
Draft the title and abstract last, then verify every number and claim against the final manuscript.
Ask a colleague outside the project to identify the first point at which the logic becomes unclear. That point is often where revision matters most.
Once the structure is solid, run the manuscript through a final source and consistency pass — our 10-minute pre-submission audit checklist covers the last-mile checks that catch exactly the kind of numbers-that-don't-travel and selective-reporting problems described above.
The Final Takeaway
A publishable manuscript is not a record of everything that happened in a project. It is a carefully structured argument that allows readers to inspect your reasoning, assess your evidence, and understand your contribution.
Your job is not to make the science sound grand. Your job is to make it clear, complete, and proportionate.
Start with a question worth answering. Describe your methods so they can be judged. Report your results without decoration. Discuss what they mean without outrunning them. Then use the right reporting guideline and the target journal's own instructions to make the manuscript submission-ready.
That is the real promise of IMRAD: not a rigid template, but a reliable way to turn hard-won research into knowledge that others can evaluate, build on, and trust.
Once your IMRAD structure holds together, use SciCampus Journal Finder to confirm the target journal's fit and quartile, then run a final pre-submission pass before you upload the manuscript.
Author Resources
International Committee of Medical Journal Editors (ICMJE). “Preparing a Manuscript for Submission to a Medical Journal.” Covers manuscript sections, including Introduction, Methods, Results, Discussion, abstracts, references, displays, and reporting guidance. https://www.icmje.org/recommendations/browse/manuscript-preparation/preparing-for-submission.html
EQUATOR Network. “How to write a great research paper using reporting guidelines.” https://www.equator-network.org/toolkits/writing-research/
EQUATOR Network. Reporting-guideline library for study designs and report sections. https://www.equator-network.org/
Related reading (topic ideas for future posts)
How to write a Methods section reviewers don't have to interrogate
Choosing the right EQUATOR reporting guideline for your study design
Abstract writing: fitting your whole argument into 250 words
What editors actually check before sending a paper out for review
