Most of the manuscripts I edit share one problem. Not grammar. Not weak data. Structure. A researcher who has spent six months in the field, who has interesting findings, loses the reader before page two because the paper is organised in the order things occurred to the writer, not in the order a reader needs to understand them.
IMRaD is the fix. It stands for Introduction, Methods, Results, and Discussion. Nearly every natural and applied science journal uses some version of this structure, because it maps the scientific method directly onto the reading experience. First: why the question matters and what nobody has answered yet. Then: exactly how it was tested. Then: what was found. Then: what it means.
Simple to state. Harder to execute well. Here is what each section is actually doing, and where most first papers fall apart.
The Hourglass Shape
Before the sections, the shape. IMRaD has an hourglass structure. The Introduction opens broad, narrows to a specific gap in knowledge, then narrows further to your specific research objective. The Methods and Results are the thin part: precise, specific, your study only. The Discussion opens back out, moving from your specific results to what they mean for the field.
The IMRaD hourglass: broad → specific → broad
Draw this shape and hold your draft against it. You will catch most structural problems before any reviewer does.
Introduction
The job of the Introduction is not to summarise the literature. It is to answer one question: why did this study need to exist?
That requires three moves. First, establish what is already known. Second, identify the gap: what is not yet known or resolved. Third, state what this study did to fill that gap. Three paragraphs, usually. Four if the field needs more groundwork.
The gap is what most introductions are missing. There is often a strong first paragraph on general background, then a direct jump to "Therefore, the objective of this study was..." with no explanation of what nobody had yet figured out. The reviewer's first question is always: why? If the gap is absent, there is no answer.
Here is the difference in practice:
The right-hand version tells the reader exactly why this study exists. The left only announces it.
One other point: the Introduction should not describe your results. "This study found that BBCU improved NUE by 15%" belongs in the Abstract or Results, not here. The Introduction states the objective, and stops.
Methods
One test determines whether your Methods are written correctly. Could another researcher at a different laboratory, reading only this section, repeat your experiment and expect comparable results? If yes, the Methods are right. If no, something is missing.
Write in past tense, passive voice (conventional in most journals, though some now accept active). Organise into subsections: study site or materials, experimental design, measurements and data collection, and statistical analysis.
That last subsection is where most first papers fail. I regularly edit manuscripts where the Results section reports ANOVA outputs and Tukey HSD comparisons, but the Methods never mention that these tests were conducted. Reviewers flag this every time. State your statistical software with version number (SPSS 26.0, R version 4.3.1), the specific tests you ran, and the significance threshold used, typically p < 0.05.
Missing statistical section entirely. No version number for software. No mention of significance threshold. Describing results ("the treatment showed higher yield") instead of procedures. Writing "data were analysed using appropriate statistical methods" — name the methods. Inappropriate passive constructions that omit what was done to what.
Results
The Results section has one job: report what happened. No interpretation, no comparison to other studies, no explanation of mechanisms. That belongs in the Discussion.
Lead with your main finding. Support it with tables and figures. Move to secondary findings. Each paragraph should point the reader to a specific table or figure and tell them what to notice: "Table 2 shows grain yield across all eight treatments. The highest yield was recorded in T4, which was significantly higher than the control (p < 0.05)."
Text should guide, not reproduce. If the numbers are in Table 2, you do not need to restate them all in the paragraph. Summarise the key pattern and let the table carry the detail.
Write in past tense throughout. "Grain yield was highest in T4" — not "Grain yield is highest in T4."
"The results are presented in Table 2." This wastes a sentence and tells the reader nothing they cannot see. Say what the table shows instead: "Table 2 shows grain yield across all treatments, with T4 producing the highest mean of 6.2 t/ha."
Discussion
This is the hardest section to write. It is also where most papers either build credibility or quietly lose it.
The Discussion answers three questions. What do these results mean? Are they consistent with what others have found, and if not, why not? What are the limitations, and what should be studied next?
Start by restating your main finding in context, without copying it word for word from Results. Then compare against the literature: "The 38% NUE observed here is consistent with values reported by [citation] in flooded paddy systems, and higher than the 31% reported by [citation] under similar conditions, possibly because the biochar matrix delayed the initial-burst release that drives most volatilisation losses." Work through secondary findings the same way.
Two mistakes to avoid. The first is a Discussion that re-summarises the Results: "Table 2 showed that T4 had the highest yield. Table 3 showed that RE was highest in T4." The reader already read the Results. The second is over-reaching: "This study proves that BBCU should replace conventional urea across Asia." A single pot experiment cannot support that claim.
"These findings suggest" is defensible. "These findings prove" rarely is. Be precise about what your data actually supports.
Conclusion
One to three paragraphs. Not a summary of the paper. The direct answer to the research question.
"This study found that BBCU, applied at 120 kg N/ha, improved agronomic efficiency by 20% and recovery efficiency by 8 percentage points compared to conventional urea in boro rice under waterlogged conditions. The biochar-bentonite matrix reduced initial-burst nitrogen release, confirmed by incubation and volatilisation trials. Local production from rice husk biochar and sodium bentonite is technically feasible and warrants field-scale validation."
Three sentences. What you found, how it works, what comes next. That is a Conclusion. It is not a re-statement of the abstract, and it does not introduce new data or references.
Abstract
Most journals print the abstract before the Introduction. Most researchers write it last. Both are correct.
The abstract follows the same IMRaD logic, compressed to 200 or 250 words. Two sentences of background, one sentence naming the gap, one sentence stating the objective, two or three sentences on methods, two or three on key results with specific numbers, and one or two on conclusions. Each sentence earns its place. There is no room for vague scene-setting.
The abstract is what most readers, including journal editors and reviewers, encounter first. Often it is the only part that gets read. Write it last, and write it with care.
References
Use a reference manager. Mendeley is free and integrates with Microsoft Word. Zotero is equally good. Do not format references by hand. Inconsistent reference formatting is one of the most common and most avoidable reasons papers are returned before peer review even starts.
Check your target journal's style before submission. Elsevier journals often use APA or Vancouver. MDPI uses a numbered system. Get it right in the template once, rather than correcting 40 entries in the manuscript after the fact.
IMRaD as a Tool, Not a Rule
IMRaD exists as a service to the reader. The reader who picks up your paper already arrives with a set of questions: why does this matter, how was it tested, what was found, what does it mean. When your structure answers those questions in that order, the science is easy to follow. When it does not, even strong results get buried.
Some journals ask for slight variations: a combined Results and Discussion section, a separate Theory section, numbered headings. These are adaptations of the same logic, not departures from it. The hourglass shape holds regardless of what the section headers say.
If you are working on a manuscript and want structural feedback before submission, the manuscript editing service on this site handles exactly this kind of work. I return tracked changes, a structural note, and a revision round within five to seven working days.
Sajjadur Rahman
MSc Researcher · Manuscript Editor · University of DhakaNST Fellow and active researcher in soil science and environmental science. Certified in scientific communication through Stanford Writing in the Sciences and the Wiley Researcher Academy. Available for manuscript editing, data analysis, thesis consultation, and IELTS tutoring.