Paper structuresPublished research
Peer-Reviewed Journal Article
The top rung: new research written for experts, checked by other experts before it's published. In the sciences it almost always follows IMRaD; in the humanities the shape follows the argument.
What it is
A report of original research, written for other researchers and published in a scholarly journal after peer review: the editor sends it to experts in the field, who judge whether the work is sound, new and worth publishing. Papers are often sent back for revision, and many are rejected.
Every paper lower on the ladder practices for this one. A school report gathers what’s known; a journal article adds something that wasn’t known, and has to show exactly how the authors know it.
In the natural and social sciences, research articles almost always use the IMRaD structure: Introduction, Methods, Results, and Discussion. It wasn’t always so. A survey of four major medical journals found no IMRaD articles in 1935; the format became the majority in the 1970s and the only pattern for original research papers in those journals by the 1980s. Humanities articles usually don’t follow IMRaD.
Who reads it
Specialists, first the reviewers and then the field. They’re short on time, skeptical by training, and looking for weak spots. Many will read only the title and abstract. A reader deciding whether to trust your result goes straight to the Methods.
So the article is written to be checked. Another researcher should be able to tell exactly what you did, judge whether your evidence supports your conclusion, and in principle do it again.
The parts
1Title, authors and abstract
The title states the finding or the question, often with the study design (“a randomized trial”), because it’s how other researchers will find the paper in a database. The abstract is a short summary of the whole article. Many journals require a structured abstract with labeled parts, such as background, methods, results and conclusions, including the main numbers.
2Introduction
What’s already known, what isn’t, and why that gap matters, ending with the specific question or hypothesis this study tests. It’s short and narrow: just enough of the literature to show where the new work fits.
3Methods
Exactly what was done: who or what was studied, how they were chosen, what was measured and how, and how the data were analyzed. Detailed enough that another team could repeat it. Fields have reporting checklists for common study types, such as CONSORT for randomized trials, which list what Methods must include.
4Results
What was found, without interpretation, usually built around tables and figures. Report effect sizes and their uncertainty, not just whether something was “significant”, and report results that didn’t support the hypothesis too.
5Discussion
What the results mean: how they answer the question, how they fit with earlier studies, what the limitations are, and what should be studied next. Good discussions don’t claim more than the results show.
6References and declarations
The references list every source cited, in the journal’s style. Around them sit the declarations journals now require: funding, the authors’ financial and other relationships that could look like conflicts of interest, each author’s contribution, and often where the data and code can be found.
What this level expects
Something new. A finding, a method, a dataset or an interpretation that the field didn’t have before.
Methods others can check and repeat. If a reviewer can’t tell what you did, the paper won’t pass.
Original sources. Cite the study itself, not a summary of it, and check that nothing you cite has been retracted.
The journal’s rules, exactly. Every journal publishes author guidelines: word limits, structure, reference style, figure formats. Papers that ignore them can be returned without review.
Full disclosure of funding and conflicts of interest.
Answering reviewers. Revision means a point-by-point response to each reviewer’s comments, saying what you changed or why you didn’t.
Common mistakes
An Introduction that reviews everything. It should lead to your question, not survey the field.
Methods too thin to repeat. “Standard procedures were used” tells a reviewer nothing.
Interpretation in the Results, or new results in the Discussion. Each section has one job.
Overclaiming. A correlation written up as a cause, or one sample’s finding written up as true of everyone.
Citing from memory or from someone else’s reference list, without reading the original. That is how errors get copied from one bibliography to the next.
Choosing a journal poorly: the wrong audience, or a predatory journal that charges fees and publishes without real review.
What varies
Nearly everything outside the core logic. Humanities articles usually have no fixed headings: they’re organized around an argument and its sub-claims, often with an introduction, thematic sections and a conclusion. Medicine adds strict reporting checklists; mathematics has theorems and proofs instead of Methods; many fields publish short letters, reviews and case reports alongside full articles. Word limits, reference styles and whether reviewers know who the authors are differ from journal to journal. Read the target journal’s author guidelines and a few recent articles it published before you start writing.
Sources & further reading
- Preparing a Manuscript for Submission to a Medical Journal — International Committee of Medical Journal Editors (ICMJE) Recommendations
- Luciana B. Sollaci and Mauricio G. Pereira, “The introduction, methods, results, and discussion (IMRAD) structure: a fifty-year survey,” Journal of the Medical Library Association, 2004
- Anatomy of a Scholarly Article — University of North Carolina at Chapel Hill Libraries
End of page · Back to top