Manuscript writing is the process of turning a completed piece of research into a document structured, formatted, and worded the way a journal’s editors, reviewers, and readers expect. The content of the work — the data, the analysis, the argument — is what makes a manuscript worth publishing, but structure and convention are what let an editor and a set of reviewers evaluate it efficiently and let readers later find the specific information they need. This guide walks through a research manuscript section by section, in roughly the order most authors actually draft them (which is not the order they appear in the finished paper), and points to CASRAI’s deeper guides on each individual section, on style, and on the submission process itself.
Most biomedical, life-science, and many social-science journals expect some variant of the IMRaD structure — Introduction, Methods, Results, and Discussion — wrapped in a title page, abstract, references, and disclosure statements. The International Committee of Medical Journal Editors (ICMJE) Recommendations set out the most widely cited reference point for what belongs in each part of a paper, and most journal-specific author guidelines are a house-style variant of this same underlying pattern. Humanities and some qualitative social-science fields structure differently (thematically rather than by IMRaD), so always check target-journal instructions before assuming this pattern applies.
Choose a Target Journal Before You Write
Word limits, section order, reference style, and even whether the abstract is structured or free-text are all set by the journal, not by a universal standard. Selecting a target journal (or a short shortlist) before drafting, and reading its author guidelines and a handful of its recently published papers in your subfield, avoids restructuring the whole manuscript later. If you’re deciding between article types — original research, review, case report, brief communication — CASRAI’s guide to journal article types covers how the expected structure changes by type.
The Title
The title is the single most-read part of the manuscript — most readers decide whether to open the abstract based on it alone, and it is what indexing systems and search engines match against a search query. It should describe what was actually studied and found, in specific, searchable terms, without unexplained abbreviations or a leading question the reader has to click through to answer. See CASRAI’s dedicated guide on writing a discoverable, well-optimized research paper title for the discoverability side of this specifically, and the related guide on choosing keywords for a research paper.
The Abstract
Most journals require either a structured abstract — separate labelled subsections, commonly Background, Methods, Results, Conclusions — or a single free-text paragraph covering the same ground in order. Either way, the abstract should stand on its own: a reader who reads nothing else should still learn the research question, what was done, the principal findings (with key numbers, not just a claim that a difference was found), and what the findings mean. Write the abstract last, once the paper’s actual results and framing are settled, even though it appears first in the manuscript.
Introduction
The introduction moves from broad context to the specific gap the study addresses, and ends by stating what the study did to address it. A common failure mode is an introduction that reads as a general literature review rather than building an argument for why this specific study was necessary. CASRAI’s guide on structuring the introduction of a research paper covers this funnel-from-context-to-gap pattern in detail with a worked structure.
Methods
The methods section needs to give a reader — and, in principle, a would-be replicator — enough detail to understand exactly what was done: study design, participants or materials, procedures, and the analytical approach, including statistical methods and software. Precision matters more than length here; vague methods language is one of the most common reasons reviewers request revision. If your study involves a sample-size or power justification, CASRAI’s guide on justifying sample size in a manuscript covers how to write that specific case in the methods section. If you’re reporting statistical results, see CASRAI’s guide on how to report p values for current formatting convention.
Results
Results report what was found, without interpreting it — interpretation belongs in the discussion. Results are typically organized around the study’s research questions or hypotheses in the same order they were introduced, using figures and tables to carry dense quantitative detail so the prose can stay focused on the findings that matter most. Every figure and table needs a self-explanatory caption; a reader skimming only the figures should be able to follow the paper’s core findings.
Discussion
The discussion interprets the results: what they mean, how they relate to prior work, what limitations bound how far the findings can be generalized, and what the implications are. It should not simply restate the results section in different words. CASRAI has two guides that go deeper on specific parts of this section: how to write the discussion section of a research paper and the limitations section of a research paper, with a worked example. For a complete worked walk-through of the full structure across all sections together, see CASRAI’s IMRaD structure worked example.
References, Figures, and Tables
Reference formatting (citation style, in-text format, reference-list order) is set by the target journal or its house style manual and should be applied consistently throughout — most reference-management software can switch between journal-specific citation styles automatically. Figures and tables should be prepared to the specific technical requirements (resolution, file format, sizing) in the journal’s author guidelines, since these are frequently checked before a manuscript is even sent for peer review.
Style Conventions
Beyond section structure, most journals expect a specific academic register: consistent tense (commonly past tense for what was done and found, present tense for established facts and the paper’s own conclusions), a defined stance on active versus passive voice, precise and hedged claims that don’t overstate what the data support, and consistent terminology and abbreviation use throughout. Two style manuals anchor much of this in practice — the AMA Manual of Style, widely used across biomedical journals, and APA Style, dominant in psychology and many social sciences — and journals that don’t specify one of these often still follow the same underlying conventions. The European Association of Science Editors (EASE) Guidelines for Authors and Translators of Scientific Articles is a useful complementary, plain-language resource, including specific guidance for researchers writing in English as an additional language.
Authorship and Disclosures
Before submission, the author list, author order, and any acknowledgments need to be settled according to the journal’s authorship criteria (most biomedical journals use or reference the ICMJE authorship criteria). Many journals now also request or require a CRediT (Contributor Roles Taxonomy) statement specifying each author’s contribution type — see CASRAI’s CRediT overview for the full taxonomy of roles. The manuscript also needs to disclose funding sources, conflicts of interest, and, increasingly, any generative-AI tool used in drafting; CASRAI’s guide on the policy landscape for generative AI in manuscripts covers what publishers currently require to be disclosed.
Before You Submit
Once the manuscript is drafted, check it against the target journal’s word-count limits (see CASRAI’s guide on manuscript word count limits and how to shorten a submission if you’re over), draft a cover letter (CASRAI has a worked, annotated cover letter example), and confirm every co-author has approved the final version. The actual mechanics of getting the manuscript into and through the journal’s submission system — the online tracking platform, editorial screening, and what happens after you submit — are covered separately in CASRAI’s manuscript submission process guide.
Frequently Asked Questions
What order should I write the sections of a manuscript in?
Most authors don’t draft in the order the sections appear in the finished paper. A common approach is methods and results first (since these are most directly tied to work already done), then the discussion, then the introduction, and the title and abstract last, once the paper’s actual findings and framing are fully settled.
How long should a manuscript be?
There’s no single universal length — word and page limits are set per journal and often per article type (original research, brief communication, case report, review), and can also vary by section (e.g., a separate abstract word limit). Always check the specific target journal’s author guidelines rather than assuming a general figure; see CASRAI’s guide on manuscript word count limits for how to handle an over-length draft.
What’s the difference between manuscript writing and manuscript submission?
Manuscript writing is drafting and structuring the paper itself — the subject of this guide. Manuscript submission is what happens afterward: uploading the finished manuscript to a journal’s editorial system, the cover letter, and the editorial screening and peer-review process that follows. CASRAI covers that stage separately in the manuscript submission process guide.
Do all journals use the IMRaD structure?
No. IMRaD (Introduction, Methods, Results, and Discussion) is the dominant pattern in biomedical and many social- and physical-science journals, and is the reference point the ICMJE Recommendations describe, but humanities journals and some qualitative social-science work use thematic rather than IMRaD structure. Review articles, case reports, and other non-original-research article types also deviate from standard IMRaD; check the target journal’s instructions for the specific article type.
Should I write the abstract first or last?
Last, in almost all cases. The abstract has to accurately summarize the manuscript’s actual results and framing, which aren’t fully settled until the rest of the paper is drafted. Writing it first often produces an abstract that no longer matches the finished paper.
