📚 Writing research articles: structure and methodology

Writing a research paper is a process that combines analytical thinking, command of academic style, and the ability to build a coherent argument. For early-career authors, the path from idea to publication can take anywhere from six months to a year and a half, and competition in top journals remains extremely tough: Nature accepts only 4-6% of manuscripts, and Science rejects about 85% at the desk rejection stage before they even reach external peer review. At the same time, the annual volume of published papers exceeded 4 million in 2024, three times more than in 2010. In this guide, we will break down the structure, methodology, and strategy that will help you get from draft to accepted manuscript in 2026.

💡 Quick overview:

• Step 1: Define your research question and formulate a hypothesis based on a literature review
• Step 2: Choose your target audience and journal before you start writing; this will determine structure and length requirements
• Step 3: Build the manuscript around the IMRaD framework (Introduction, Methods, Results, Discussion), the standard for most disciplines
• Step 4: Prepare a cover letter with a clear statement of the scientific contribution
• Step 5: Check the manuscript against the journal's requirements and submit; if rejected, revise and submit to the next journal

The IMRaD structure: a framework editors understand

The IMRaD format is an international standard for scientific papers, recognized by more than 2,000 biomedical journals and recommended by the International Committee of Medical Journal Editors (ICMJE). The acronym stands for Introduction, Methods, Results, and Discussion. This structure is not just a formality: it mirrors the logic of scientific inquiry and helps reviewers quickly assess the soundness of the work.

In the social sciences and humanities, the structure may vary: instead of Methods, authors use a theoretical framework, and the Results section is often combined with Discussion. However, the principle of "introduction, main body, conclusions" is universal. As Trinka AI methodologists note, editors value manuscripts where each section does its own job without overlap.

The English-language video above gives a clear 5-minute breakdown of the IMRaD format; watch it before you sit down to draft. After the video, we will move on to journal selection strategy: even a brilliant manuscript will not survive desk rejection if it is submitted to the wrong place.

How to choose a journal: desk rejection and other barriers

Desk rejection, when an editor turns down a manuscript before sending it to reviewers, cuts out 30 to 90% of submissions depending on the journal. At the New England Journal of Medicine this threshold reaches 90%, at Science, 85%, and at Nature, about 60%. The main reason is a mismatch with the journal's scope and level, not the quality of the work. A Nature editor has 5-7 business days to assess whether a manuscript has broad interdisciplinary appeal; if not, it is rejected, and the authors lose a week when they could have submitted directly to a specialized journal.

The practical takeaway: submitting to a top journal is justified only if the work has interdisciplinary significance. Otherwise, the average paper goes through 2.5 journals before publication, and 40% of researchers get rejected by three or more outlets. This is normal: Charles Darwin and Albert Einstein both faced rejections from leading journals. Choose a journal based on the principle of "one target specialized journal plus one backup," not "Nature or nothing."

Behind the scenes of peer review: timelines, numbers, and reality

After a manuscript passes the desk stage, it goes to reviewers, and the waiting begins. According to a survey of 3,500 authors on the SciRev platform, the average time to a first decision is 13 weeks, ranging from 8 weeks in medicine to 18 weeks in economics and business. Only 19% of authors receive a response within a month, and 32% wait three months or longer. In 2026, a study of 57 health policy journals recorded a median of 60.5 days to a first decision, with a range of 21 to 263 days.

Globally, the research community spends about 68.5 million hours per year on peer review, equivalent to 7,800 researcher-years. At the same time, authors collectively spend 260 million researcher-days waiting for reviews, roughly 720,000 person-years. The number of invitations needed for one completed review has grown from 1.9 in 2013 to a projected 3.6 in 2025: reviewers are in short supply, and publishers compete for their time.

What this means for authors: submit the manuscript and immediately start working on the next one. Waiting six months for a single response is strategic downtime, especially for PhD students and grant applicants.

Working with revisions: revise and resubmit as the norm

Roughly 25-30% of manuscripts at mid-tier journals receive a "revise and resubmit" decision in the first round of review. This is good news: the editor sees potential. The bad news is that authors have to respond to every reviewer comment, often on a tight deadline.

Practical rules for handling revisions:

• Respond to each comment as a separate item in the cover letter. Do not ignore any comment, even if you disagree.
• If you reject a reviewer's suggestion, explain why, citing data or methodology rather than personal opinion.
• Keep two versions of the manuscript: one with tracked changes and one clean copy. Most editors ask for both.
• Meet the deadline set by the editor. Missing it lowers the chances of final acceptance.
• Check whether the reviewers' comments contradict each other. If they do, ask the editor for arbitration before revising the text.

Rejection after peer review is not a death sentence. A 2023 study of 4,000 researchers showed that manuscripts rejected by one journal after peer review have comparable chances of publication in a journal of the same tier, provided the comments are addressed. Revise, polish, and submit elsewhere.

Tools and open access: the realities of 2026

The technology landscape of academic writing has changed noticeably. According to 2025 estimates, 15 to 20% of manuscripts in leading journals contain AI-assisted text. Nature, Science, and Elsevier journals now require explicit disclosure of AI use in manuscript preparation. AI cannot be a co-author, but editors acknowledge its role in literature search, proofreading, and data analysis.

Open access (open access, OA) covers about half of new articles in 2026. The Plan S initiative, backed by European funders (including UKRI), requires immediate OA for funded research. Publication fees (APC) range from about 500 dollars to 11,590 dollars at Nature. At the same time, OA articles receive on average 25-30% more citations than paywalled articles, when controlling for journal level.

The practical minimum set of tools for an author in 2026:

• Reference manager: Zotero (free, open source) or EndNote. Without a reference manager, formatting the bibliography to meet a journal's requirements turns into hours of routine work.
• Preprint server: arXiv, bioRxiv, SSRN, or PsyArXiv. In 2024, over 2.5 million preprints were posted on the major servers, roughly 60% of all published peer-reviewed articles in those fields.
• Plagiarism check: iThenticate or Turnitin. Most journals run manuscripts through these automatically; checking before submission removes risks.
• AI writing assistant: tools like Trinka, Paperpal, or Writefull for language editing, especially relevant for authors who are not native English speakers.

An important caveat: no AI tool guarantees scientific soundness. Reviewers evaluate the argument, not the style, and replacing thinking with text generation leads to superficial manuscripts that desk rejection filters out in minutes.

⁉️🤔 Frequently asked questions

Is it true that an academic career is impossible without Scopus/WoS publications?

Requirements depend on the country and discipline. At research universities in the US and Europe, a successful candidate for a tenure-track position has an average of 8-12 peer-reviewed publications. At the same time, 73% of early-career researchers report significant pressure to publish, and 89% of search committees name productivity as the most important criterion. At British universities, the average academic produces 2-3 articles per year. However, in the humanities a monograph often carries more weight than a series of articles, and in engineering conference proceedings may be the priority.

How long does publication take from submission to publication?

The median time from submission to first decision is 37 days. After acceptance, another 4 to 12 months pass before online publication. So the full cycle rarely fits within a year. The process is fastest in medicine (about 8 weeks to first response) and slowest in economics (18 weeks). Open access journals with APCs publish faster: 4-6 months after acceptance.

What should I do if reviewers contradict each other?

This is a common situation: the correlation between two reviewers' assessments of the same manuscript is only 0.34 according to a meta-analysis of 45 studies. If the comments are incompatible, write to the editor before starting revisions: lay out the contradiction and ask them to indicate the priority. The editor will either resolve the conflict or bring in a third reviewer. Do not try to satisfy everyone at once: the text will become internally inconsistent.

Is open access publishing mandatory?

Not mandatory, but desirable. About 50% of new articles are already OA, and that share is growing under pressure from funder mandates. The main argument: OA articles are cited 25-30% more often. If a grant or institution requires OA, check whether the grant covers the APC. For researchers without funding, the number of "diamond" OA journals is growing: publication is free for both the author and the reader, especially in the humanities.

How do you tell a predatory journal from a legitimate OA outlet?

Three quick signs: the journal promises publication within days (real peer review takes weeks or months), sends spam invitations by email, and is absent from DOAJ (Directory of Open Access Journals) or Web of Science. Use the Think-Check-Submit checklist and check against your institution's "white lists." A predatory journal will not just take your APC: publishing in one can damage your reputation in future submissions.

How many articles does a PhD student need to publish?

In STEM disciplines at research universities, a PhD student is expected to produce 2-4 peer-reviewed publications during the program. In the humanities and social sciences the bar is lower: 1-2 journal articles or conference papers, and the dissertation itself is considered the main publication. Check your department's requirements in the first year: in some programs, publications are mandatory for admission to the defense.

Summary

A research article in 2026 is a product that passes through a multi-stage filter: journal selection, desk rejection, external peer review, revisions, and the editor's decision. The statistics are sobering: up to 90% of manuscripts submitted to top journals do not pass the desk, the average article changes 2.5 outlets before publication, and the cycle from submission to publication takes a year. But the same statistics suggest a strategy: choose the journal before writing, structure the text according to the IMRaD standard, invest time in responses to reviewers, and use modern tools, from reference managers to preprints. Academic publishing stops being a lottery when it is treated as a reproducible process with its own rules. Start with a draft today, and in six months your ORCID will have a new entry.