The IMRaD Structure
Almost every empirical paper (a lab report, a journal article, your thesis) is built from the same four blocks: Introduction, Methods, Results, and Discussion. The four blocks come in the order a reader needs to follow your reasoning. Once you know what each section is for (each has one job), most writing decisions ("where does this sentence go?") become easy.
Four sections, one job each
The whole structure fits in four sentences. If a paragraph you're writing doesn't do the job below, it's in the wrong section:
- Introduction: why the study. Set up the problem, review what's known, find the gap, and end on your specific hypothesis or question. It moves from broad context to one narrow prediction (this is where your H₀ and H₁ come from).
- Methods: what you did. Who took part, what you measured, and how you analyzed it, in enough detail that a stranger could repeat the study. No results, no interpretation; just the recipe.
- Results: what you found. The descriptives and test statistics, reported with no spin. Numbers and what they say directly, not yet what they mean.
- Discussion: what it means. Interpret the findings, connect them back to the literature, admit the limitations that matter, and point to what's next. This is the only section where you're allowed to editorialize.
The hourglass
IMRaD has a shape: an hourglass. The Introduction starts wide (the big topic everyone cares about) and narrows to your exact hypothesis. The Methods and Results are the tight neck: specific, concrete, about your study only. Then the Discussion widens back out, from your particular finding to what it means for the field.
Reading order ≠ writing order. Readers go top to bottom, but you shouldn't write that way. Draft the Methods and Results first. They're the most factual and the easiest to pin down. Then write the Introduction (now you know exactly what you're introducing) and the Discussion. Write the abstract last, once everything it summarizes exists.
How much detail is "enough detail"?
"Enough that a stranger could repeat the study" is a general standard, and students often find it hard to apply. APA publishes a detailed version. Since the 7th edition of the Publication Manual, the Journal Article Reporting Standards (JARS) list what each section has to contain, in tables free to read at apastyle.org/jars. There are four sets: JARS-Quant for numerical data, JARS-Qual for natural-language data, JARS-Mixed for both, and JARS-REC for reporting race, ethnicity, and culture in any of them.
Most theses split the Methods into four subsections (Participants, Materials or Measures, Procedure, Data Analysis). JARS is the list of what has to appear under them, and it is longer than most first drafts:
- Inclusion and exclusion criteria, then who actually took part: age, sex, ethnicity, and anything else your topic makes relevant.
- How participants were selected, where and when data were collected, what they were paid, and which ethics committee approved it (see Sampling Methods and Informed Consent & Ethics Committees).
- Sample size, power, and precision: the size you intended, the size you achieved, and how you decided on the first one (a power analysis, or the power calculator).
- Every measure and covariate you collected, including the ones that never made it into the report. Most drafts miss this item, and it is there on purpose: if you list the three things you measured but never analyzed, a reader can judge how many analyses were available to you.
- The design and conditions (between or within), and who was kept unaware of them (Bias & Blinding).
- The analytic strategy: which tests, how missing data were handled and how much there was, and what protected you against inflated error rates across multiple tests.
Read that list before you draft, not after, and you won't have to write the Methods from memory. JARS also asks the abstract to report effect sizes and confidence intervals, not just significance. That is why the next lesson, Reporting Statistics in APA Style, spends as long on the interval as on the p-value.
The classic misplacements
Graders see the same four mistakes every year, and every one is a section doing another section's job:
- Interpretation in the Results. "This shows that napping boosts memory" is a claim, and it belongs in the Discussion. Results just report the number.
- Findings in the Discussion for the first time. Every result you interpret must have already appeared, with its statistics, in the Results section.
- New literature in the conclusion. If a citation matters, introduce it in the Introduction. Don't spring fresh sources at the end.
- Method details in the Results. "We used a 40-item recall task" describes procedure, so it belongs in the Methods, not next to the numbers.
Sort the sentences
Below are twelve sentences from one finished study, a nap-and-memory experiment, in jumbled order. File each into the section it belongs to. If you miss one, you'll see what that section's job is, so you can work it out.
🗂️ Which IMRaD section does each sentence belong to?
Click a section to file each sentence. Green = filed correctly; a miss explains what that section is for.
Read in order, the sections make the argument. The four sections give the whole argument in brief: the question (I), how I looked (M), what I saw (R), what it means (D). If a reader gets lost, a sentence is usually in the wrong section, and moving it fixes the problem.
Why it matters: IMRaD tells your reader where to find things. If you follow it, a busy examiner can go straight to your Methods to check your design, or to your Results to check your numbers. If you bury a finding in the Discussion or put interpretation in the Results, the reader has to do the sorting that you should have done.
Common questions
What does IMRaD stand for?
IMRaD is the standard structure of an empirical paper: Introduction, Methods, Results, and Discussion. Each section has one job. The Introduction says why the study (context → gap → hypothesis). The Methods say what you did (in enough detail to replicate). The Results say what you found (the numbers, with no interpretation). The Discussion says what it means (interpretation, limitations, and what's next). If a paragraph doesn't do the job of the section it's in, it's in the wrong place. The lowercase 'a' is just 'and'.
Should I write a paper in the order it's read?
Usually not. Trying to write in reading order is a common reason students get stuck on a blank page at the Introduction. Draft the Methods and Results first, because they're the most factual and concrete: you already know what you did and what you found. Then write the Introduction (now you know precisely what you're setting up) and the Discussion (now you know what you're interpreting). The hypothesis in your Intro should match the analyses in your Results, which is far easier to guarantee if the Results already exist. Write the abstract last, once everything it summarizes is finished.
Where do I put the interpretation of my results?
In the Discussion, never in the Results. The Results section reports the finding without comment ("the nap group recalled more word pairs, t(58) = 3.72, p < .001") and stops there. Saying what that finding means ("this supports the idea that sleep aids consolidation"), how it fits prior work, or why it might have happened is interpretation, and interpretation belongs in the Discussion. The mirror-image mistake is just as common: don't introduce a finding for the first time in the Discussion. Everything you interpret there must already have appeared, with its statistics, in the Results.