A scientific paper follows a standardized structure designed to let readers quickly find what they need. Most journals use what’s known as the IMRaD format: Introduction, Methods, Results, and Discussion. Each section has a specific job, and understanding those jobs before you start writing will save you dozens of revision hours. Here’s how to work through each part, from first draft to submission.
Start With the Structure, Not the Writing
Readers of scientific papers rarely read front to back. They jump to the section that answers their immediate question: a clinician might check the results first, while a fellow researcher might go straight to the methods. The IMRaD structure exists to support this kind of modular reading. Before you write a single paragraph, outline your paper into these core sections: Abstract, Introduction, Methods, Results, Discussion. Some journals add a Conclusion as a separate section, and many require additional elements like a data availability statement or conflict of interest disclosure.
A practical approach is to write the sections out of order. Many experienced researchers draft the Methods and Results first (since those are the most concrete), then the Discussion, then the Introduction, and finally the Abstract. This prevents you from writing an introduction that promises something your data doesn’t deliver.
Writing the Introduction
The introduction answers three questions: What do we already know about this topic? What don’t we know? How does your study address that gap? The most effective way to organize this is the “funnel” approach. You start broad, establishing the general topic and why it matters, then gradually narrow the focus through what’s currently known in the literature, what remains unanswered, and finally your specific research question or hypothesis.
Think of it as five narrowing layers: general topic, specific background, the knowledge gap, your study’s purpose, and your hypothesis or aims. Each layer should flow logically into the next, so the reader arrives at your research question feeling like it was the obvious next step. Keep it concise. Most introductions run three to five paragraphs. The goal is context, not a comprehensive literature review.
The Methods Section
The methods section has one overriding purpose: another researcher should be able to repeat your experiment using only what you’ve written here. That standard of reproducibility determines what to include. Describe your study design, all materials you used, the procedures you followed, how you measured your outcomes, and how you analyzed the data.
For materials, be specific. Include equipment names, model numbers, and manufacturers. If you used chemicals, drugs, or biological reagents, list concentrations and sources. If your equipment has adjustable settings, state the exact settings you used. These details might feel tedious, but they’re often the difference between a reproducible study and one that can’t be verified.
For statistical analysis, name the tests you ran and the software (including version number) you used. If you set a significance threshold, state it. If you handled missing data in a particular way, explain that too. Write the methods in past tense and in enough detail that a colleague in your field could follow the steps without emailing you for clarification.
Presenting Your Results
The results section reports what you found, without interpreting why. Present your data in a logical sequence that mirrors the order of your methods. Lead with your primary outcome, then move through secondary findings.
When reporting statistics, go beyond simple p-values. Current best practice is to report effect sizes (like the difference in means between groups, or standardized measures like Cohen’s d) alongside 95% confidence intervals. A confidence interval tells readers the range of plausible values for your finding, which communicates much more than a binary “significant or not.” If you do report p-values, state the exact value (p = 0.03, not just p < 0.05), define whether it’s a one-tailed or two-tailed test, and always pair it with an effect size so readers can judge the practical significance of your finding, not just the statistical significance.
Avoid duplicating the same data in both a figure and a table. Use tables for precise numbers that readers might want to look up individually, and figures for patterns, trends, or comparisons that are easier to grasp visually.
Figures and Tables That Meet Journal Standards
Journals are strict about image quality. The minimum resolution for most figures is 300 DPI (dots per inch). If your figure combines photographs with text or line art, aim for 600 DPI. Pure black-and-white line art, like flow charts or diagrams, should be 1,000 DPI. Preferred file formats are typically TIFF, EPS, or AI. Avoid saving figures as JPEGs, since JPEG compression degrades image quality. If file size is a concern, use lossless compression like LZW instead.
Every figure and table needs a clear, self-contained caption. A reader should understand what the figure shows without reading the main text. Label axes, include units, and define any abbreviations directly in the caption.
Writing the Discussion
The discussion is where you explain what your results mean. Open with a short paragraph summarizing your key findings and reminding the reader why the study matters. Then spend three to four paragraphs comparing your results to existing literature, starting with your most important finding and working toward secondary ones.
A common mistake is only citing studies that agree with your results. If your findings contradict previous work, say so, and offer possible explanations. Differences in methodology, sample size, or study population are all legitimate reasons results might diverge. Unexpected or negative findings deserve attention too, since they often point toward the most interesting open questions in a field.
Dedicate at least one paragraph to limitations. Be honest about potential sources of bias, any confounding variables you couldn’t control for, and the role that chance might have played. For each limitation, indicate the direction it might have pushed your results (toward or away from a significant finding). Reviewers and editors look for this kind of intellectual honesty, and papers that skip it are more likely to be rejected.
Close the discussion with a paragraph on implications: how your findings advance the field, what new questions they raise, and what practical applications they might have.
The Abstract Comes Last
Your abstract is a self-contained summary of the entire paper, typically 150 to 250 words (about six or seven sentences). Many journals require a structured abstract with labeled subsections like Background, Methods, Results, and Conclusions. Others accept a single unstructured paragraph. Always check your target journal’s specific requirements before writing this section.
Write the abstract after everything else is finished. It should include your research question, a brief description of your methods, the main results with key numbers, and your primary conclusion. Avoid vague statements like “results are discussed.” Instead, state the actual finding. The abstract is the most widely read part of any paper, and for many readers, it’s the only part they’ll see.
Managing Citations Efficiently
Citation formatting is one of the most tedious parts of manuscript preparation, and switching between citation styles (APA, Vancouver, Chicago) manually is a recipe for errors. Reference management tools handle this automatically. The three most widely used are Zotero (free, developed by George Mason University), Mendeley (free basic version, owned by Elsevier), and EndNote (free online version, desktop version costs $128 to $275). All three let you save references from databases, organize them into folders, and insert formatted citations directly into your word processor. Pick one early in your research career and use it consistently.
Authorship and Ethics
Before you submit, make sure your author list follows established guidelines. The International Committee of Medical Journal Editors requires that every listed author meet all four of these criteria: they made substantial contributions to the design or data analysis; they helped draft or critically revise the manuscript; they approved the final version; and they agreed to be accountable for the work’s integrity. Someone who only provided funding or collected data without meeting the other criteria should be acknowledged, not listed as an author. Getting this wrong can create serious professional conflicts, so have the authorship conversation early.
What Happens After You Submit
Once you submit to a journal, your manuscript enters peer review. In medical, public health, and natural science journals, the total review time averages 12 to 14 weeks. In fields like economics and business, expect 25 weeks or more. In extreme cases, the process can take over a year. You’ll typically receive one of four decisions: accept, minor revisions, major revisions, or reject.
Most papers are not accepted on the first submission. A request for revisions is normal and often a good sign. When responding, address every reviewer comment point by point in a separate response letter, explaining what you changed and why. If you disagree with a reviewer’s suggestion, provide a clear, evidence-based rationale. Editors appreciate thorough, respectful responses, and a well-handled revision can turn a borderline paper into a published one.