Why the shape exists
IMRaD became the default because it matches the logic of empirical argument: here is the question, here is how I tried to answer it, here is what I observed, here is what I think it means. Separating observation from interpretation lets a reader evaluate the evidence before being told what to conclude from it, which is the whole point of the arrangement. Almost everything else about the format follows from that one separation.
The format also makes papers navigable. A reader who wants only the method, or only the findings, knows exactly where to go, and that predictability is valuable at the scale modern literatures operate on. Journals reinforce it because it makes peer review faster and comparison across papers straightforward.
It is worth understanding this rationale rather than treating IMRaD as an arbitrary template. Most section-boundary questions — where does this sentence go? — resolve immediately once you ask which of the four jobs the sentence is doing.
What each of the four parts owns
Each section answers one question and should contain nothing that answers a different one. The boundaries are strict by design, and most structural problems in a draft are a sentence sitting in the wrong section rather than a missing idea. Moving one paragraph to where it belongs often fixes what felt like a much larger problem.
A quick test for any paragraph: could a reader holding your dataset verify it? Then it belongs in the results. Could a reader repeat your study from it? That is methods. Anything arguing why the question was worth asking is introduction, and anything saying what the findings mean is discussion.
- Introduction — why was this question asked? Context, gap, aim, and hypothesis.
- Methods — what was done? Design, participants or materials, procedure, analysis, ethics.
- Results — what was found? Data and patterns, with no explanation of them.
- Discussion — what does it mean? Interpretation, comparison, limitations, implications.
The hourglass
The four sections form a recognizable shape. The introduction starts broad and narrows to a specific question; the methods and results are the narrow waist, dealing only with your study; the discussion widens again from your findings back out to the field. Drawing this on a page is a fast way to diagnose a draft.
The symmetry is a useful check. If your introduction opens on a large claim about the field, your discussion should return to it and say what your findings do to that claim. If it never gets back there, either the introduction opened too wide or the discussion stopped too early.
The waist should be genuinely narrow. Methods and results that keep referring outward — to other studies, to what the finding might mean — have widened where the shape calls for restraint. Save that material; it is the raw material of a stronger discussion.
What sits around the four sections
IMRaD names the body of the paper, not everything in it. Before the introduction come the title, abstract, and keywords; after the discussion come acknowledgments, funding and conflict-of-interest statements, references, and any appendices or supplementary material. Journals increasingly collect the declarations through the submission system rather than in the manuscript.
Several optional sections attach to the four. A separate Conclusion is common, though many journals fold it into the discussion. Limitations sometimes get their own subheading. Some fields place a short Literature Review after the introduction, and some journals ask for a Theoretical Framework section, particularly in education and management.
Two variants are worth knowing. IMRaD with a combined Results and Discussion appears where each finding needs immediate context, common in qualitative and case-study work. IMRaD with methods placed last is used by a few high-profile journals, which changes reading order but not the logic of the sections.
Where IMRaD does not fit
The format assumes data collected through a procedure. Papers that argue rather than report — theoretical contributions, conceptual analyses, philosophical work, legal scholarship, and most humanities articles — are organized by the structure of the argument instead, with headings that name the claims being made. Trying to fit them into IMRaD produces an empty methods section and a results section that is really analysis.
Review articles have their own shapes. Narrative reviews are organized thematically; systematic reviews follow reporting guidelines that specify their own structure, including search strategy, screening, and synthesis, and are close to IMRaD without being identical. Meta-analyses follow the same guidelines with an added statistical layer.
Theses are related but not the same. A monograph thesis usually has an IMRaD-like spine — introduction, literature review, methodology, findings, discussion, conclusion — spread across chapters, while a thesis by publication wraps several IMRaD papers inside an introductory and concluding frame. Check your institution’s regulations, because these are formally specified more often than people expect.
