Nagarjun's blog: What Is a Scientific Paper? A Field Guide to the Strange Little Creature Called “the Literature”

A scientific paper is not just a document. It is a claim-making machine.

At its best, it says:

“Here is a question. Here is what we did. Here is what we found. Here is how confident you should be. Here is what this does not prove. Here is enough information for others to inspect, repeat, challenge, or build on it.”

That is the soul of the scientific paper. Not the PDF. Not the DOI. Not the fancy journal logo perched like a jeweled beetle on the first page. A scientific paper is a structured contribution to the shared memory of science.

But the moment we say “paper,” the jungle thickens. There are original articles, reviews, systematic reviews, meta-analyses, brief communications, letters, case reports, conference papers, preprints, protocols, data papers, registered reports, editorials, corrections, and database records. Some are peer-reviewed. Some are not. Some count as prior publication. Some usually do not. Some are indexed. Some are merely discoverable. Some are real papers wearing informal clothes. Some are just abstracts with ambition.

Let us sort the cabinet 🧪📚.

1. The scientific paper as a public claim

A scientific paper is a formal scholarly work that communicates research, analysis, interpretation, or commentary to a community. In modern journal publishing, the most important version is usually the Version of Record, the fixed, citable version formally published by a journal or publisher. DOI infrastructure such as Crossref helps make research objects persistent and discoverable by registering metadata and persistent identifiers, but a DOI itself does not magically certify quality or peer review. It is a signpost, not a halo. (www.crossref.org)

A good scientific paper does three jobs at once:

Job	What it gives the reader
Communication	What was done and found
Certification	Whether the work passed some editorial or peer-review process
Archiving	A durable, citable record for future scholarship

The third job is often overlooked. A paper is not only written for today’s reader. It is written for a future graduate student, a systematic reviewer, a patent examiner, a meta-analyst, a policy writer, a skeptical rival, and some tired scientist in 2037 trying to understand why Figure 2B mattered.

2. Primary literature: where new evidence first enters the room

Primary literature reports original research or first-hand scientific evidence. In the sciences, primary literature usually includes original research articles where authors describe their own experiments, observations, fieldwork, clinical data, simulations, datasets, or analyses. Library and information-science guides commonly define primary scientific literature as work that presents original research or new scientific discoveries. (library.onu.edu)

Examples of primary literature include:

Type	What it contains
Original research article	New experimental, observational, clinical, computational, or theoretical findings
Clinical trial report	Results of a study where participants are assigned interventions according to a protocol
Case report	Detailed report of an unusual clinical case or small case series
Brief communication	Short report of an important but compact finding
Methods paper	New experimental, analytical, computational, or statistical method
Data paper	Description of a dataset, usually with metadata and reuse guidance
Registered report	Study accepted in principle before results are known, usually after peer review of rationale and methods

The National Library of Medicine’s publication types include categories such as clinical trial, review, case reports, systematic review, randomized controlled trial, and others, showing how databases classify scholarly works by their content or style. (National Library of Medicine)

A primary paper is where the evidence first walks onstage.

3. Secondary and tertiary literature: the evidence gets digested

If primary literature is the raw harvest, secondary literature is the kitchen.

Secondary literature analyzes, interprets, evaluates, or synthesizes primary studies. Review articles are the classic example. An NCBI Bookshelf chapter describes review articles as journal-length papers whose purpose is to synthesize literature in a field, without collecting or analyzing new primary data. (NCBI)

Examples:

Type	Purpose
Narrative review	Broad expert synthesis, often less protocolized
Systematic review	Structured search, selection, appraisal, and synthesis of evidence
Meta-analysis	Statistical pooling of results from multiple studies
Scoping review	Maps the range and nature of evidence in a field
Perspective or opinion	Interpretive argument, often expert-led
Editorial	Journal or invited commentary on a paper, issue, or field

Tertiary literature includes textbooks, encyclopedias, manuals, and reference works. These usually summarize established knowledge rather than introduce new evidence.

A useful shortcut:

Primary literature says: “We found this.”
Secondary literature says: “Here is what many studies together suggest.”
Tertiary literature says: “Here is what the field currently teaches.”

4. Not every scientific paper has the same job

Scientific papers are like tools in a surgical tray. Mistaking one for another leads to scholarly finger injuries.

Original research article

This is the standard primary research paper. It usually contains Introduction, Methods, Results, and Discussion. The ICMJE notes that original research articles are often organized in this structure because it reflects the process of scientific discovery. (ICMJE)

Short communication or letter

This reports a compact but important observation. It may have fewer figures, shorter methods, and limited discussion. It is still primary literature if it reports new data.

Case report

A case report describes one or a small number of patients. It is valuable for rare presentations, unexpected adverse effects, new syndromes, or clinical teaching. It usually sits lower in evidence hierarchies than randomized trials or systematic reviews, but it can be the first flare in the fog.

Methods paper

A methods paper introduces or validates a technique. It can be primary literature if it presents original performance data, benchmarking, validation, or implementation.

Protocol paper

A protocol paper describes what will be done before results are available. It is not a results paper. Its value lies in transparency, preventing outcome-switching, and helping others inspect the planned design.

Registered report

A registered report flips the usual publishing suspense. The rationale and methods are peer-reviewed before data collection or before results are known. If the study is conducted as approved, the journal commits to publication regardless of whether results are positive, negative, or inconvenient.

Data paper

A data paper describes a dataset: how it was generated, curated, validated, formatted, and how it can be reused. The dataset itself may live in a repository, but the data paper is a citable scholarly article about the dataset.

Review article

A review article synthesizes what is already known. A systematic review or meta-analysis can be extremely important, but it is usually secondary literature because it analyzes prior studies rather than reporting new experimental data.

Editorial, commentary, correspondence

These are scholarly conversation pieces. They may critique, contextualize, praise, question, or extend published work. They are part of the literature, but they are usually not primary evidence unless they contain new data.

Correction, expression of concern, retraction

These are maintenance tools for the scholarly record. They are not glamorous, but they are essential. Science without correction would become a museum of polished mistakes.

5. Abstracting and indexing services: the librarians of the literature

Publishing and indexing are not the same thing.

A journal publishes a paper. An indexing or abstracting service makes that paper findable, classifiable, searchable, and sometimes measurable.

Important examples:

Service	What it does
PubMed	Search platform for biomedical citations from MEDLINE, life science journals, and online books
MEDLINE	Major curated component of PubMed, with selected journals and MeSH indexing
PubMed Central	Free full-text archive of biomedical and life-science journal literature
Scopus	Abstract and citation database curated by independent subject experts
Web of Science Core Collection	Curated multidisciplinary citation index
DOAJ	Index of open-access journals committed to quality open availability
Crossref	DOI and metadata infrastructure linking scholarly records

PubMed comprises more than 40 million citations from MEDLINE, life science journals, and online books, while PubMed Central is a free full-text archive maintained by the U.S. National Library of Medicine. (PubMed) MEDLINE is the largest component of PubMed and includes journals selected for MEDLINE, with MeSH indexing and curated metadata. (PubMed)

This distinction matters. A paper having a PMID means it has a PubMed record. It does not automatically mean the journal is indexed in MEDLINE. The NLM explicitly decides whether a journal’s scientific and editorial quality merit MEDLINE inclusion. (National Library of Medicine)

Scopus describes itself as a source-neutral abstract and citation database curated by an independent Content Selection Advisory Board, while Clarivate describes Web of Science Core Collection as a curated multidisciplinary citation database built through editorial selection. (www.elsevier.com) DOAJ is an index of open-access journals that aims to ensure quality content is openly available. (Directory of Open Access Journals)

A tiny but important warning:

Indexed does not mean perfect.
Not indexed does not always mean bad.
But indexing does affect visibility, credibility, evaluation, and discoverability.

A paper can exist without being indexed. But without indexing, it may drift like a message in a bottle across a very large ocean.

6. Preprints: papers before the ceremonial gate

A preprint is an author’s version of a research manuscript made publicly available before formal peer review by a journal. Springer Nature defines preprints as author versions of research manuscripts deposited on public servers before formal peer review. (Springer Nature Support)

Preprints are important because they:

make work visible quickly,
establish a timestamp of priority,
invite community feedback,
allow early sharing during fast-moving research,
improve access for readers without subscriptions.

But a preprint is not the same as a peer-reviewed journal article. It is a manuscript in public, not a final certified record. It may later change substantially, be rejected, be published elsewhere, or never be formally published.

Many major publishers do not treat preprints as prior publication. Nature Portfolio says posting preprints is not considered prior publication and will not jeopardize consideration at its journals. (Nature) Springer Nature similarly states that posting preprints is not considered prior publication for Springer Nature journals. (Springer Nature Support) ACS also allows authors to deposit initial drafts in preprint services such as ChemRxiv, bioRxiv, and arXiv. (American Chemical Society Publications)

But the rules are not universal. COPE emphasizes that what counts as prior publication varies between journals and disciplines, and that journals should communicate their policies clearly. (Publication Ethics)

So the practical rule is:

A preprint is usually not prior journal publication, but always check the target journal’s policy.

Preprints are the scholarly equivalent of opening the lab window before the official lecture begins. Useful, democratic, breezy, occasionally chaotic.

7. What counts as prior publication?

This is where the swamp bubbles.

There is no single global police officer of prior publication. The answer depends on journal policy, publisher policy, discipline, article type, extent of overlap, copyright status, peer-review status, and transparency.

The ICMJE gives influential biomedical guidance on overlapping publications. It says authors should not submit the same manuscript to more than one journal simultaneously, because this can waste peer-review effort and create conflicts between journals. (ICMJE) ICMJE also recognizes acceptable secondary publication under certain conditions and treats manuscripts based on the same database as potentially independent if they differ in analytical methods, conclusions, or both. (Research Integrity in Law)

Here is a practical map:

Material already public	Usually considered prior publication?	What to do
Same full paper already published in a journal	Yes	Do not submit as original article
Same manuscript under review elsewhere	Duplicate submission, not allowed	Submit to only one journal at a time
Preprint on recognized server	Usually no, but journal-specific	Disclose and cite/link it
Conference abstract	Often no, if brief	Disclose if relevant
Full conference proceedings paper	Sometimes yes, depending on overlap and field	Check journal policy, cite, explain extension
Thesis or dissertation in repository	Often no, but journal-specific	Check policy, disclose if asked
Poster or oral presentation	Usually no	Mention if required
Protocol registration	No, usually required or encouraged	Provide registry details
Dataset in public repository	Usually no, but it is public data	Cite accession/DOI and disclose
Previously published figures or text	Yes for that material	Get permission, cite, avoid duplicate publication
Pre-registered report or protocol paper	It is a publication, but results paper may still be valid	Cite and distinguish results
Blog post or non-peer-reviewed essay with same analysis	Depends on content and journal	Disclose, avoid text recycling

For theses, COPE does not impose one universal answer. It says journals and publishers need clear guidance for authors on prior publication and theses. (Publication Ethics) For conference proceedings, Nature Portfolio policies ask authors to disclose proceedings papers, cite them, obtain permissions for reused material, and attribute appropriately. (Nature)

The hidden monster here is not merely “prior publication.” It is redundant publication, text recycling, and salami slicing.

Redundant publication means publishing substantially the same work more than once without proper justification.
Text recycling means reusing your own text without transparency or permission where needed.
Salami slicing means splitting one study into multiple thin papers without a valid scientific reason.

A second paper from the same dataset can be legitimate if it asks a different question, uses different analyses, and clearly cites the related work. ICMJE notes that manuscripts based on the same database may be considered independently when analytic methods or conclusions differ. (Research Integrity in Law)

8. Trial registration is not publication, but it is public accountability

Clinical trial registration is not the same as publishing a paper. It is a public record of a planned or ongoing study. ICMJE says the purpose of clinical trial registration is to prevent selective publication and selective reporting, prevent unnecessary duplication, inform patients and the public, and help ethics boards see related work. (ICMJE)

This is crucial. A trial registry entry tells the world:

“This study exists, these outcomes were planned, and the results cannot quietly vanish if inconvenient.”

A trial registry is not a paper, but for clinical research it is part of the ethical scaffolding around the paper.

9. Is depositing in a database “publishing”?

This is the excellent, thorny question.

The answer is: it depends what you mean by publishing.

In the broad sense

Yes, depositing data in a public database is a form of public dissemination. It creates a timestamped, citable, accessible research object. GenBank, for example, is the NIH genetic sequence database and an annotated collection of publicly available DNA sequences. It is part of the International Nucleotide Sequence Database Collaboration, along with DDBJ and ENA, and these organizations exchange data daily. (NCBI)

In the journal-publication sense

No, a database deposit is usually not equivalent to a peer-reviewed research article. A GenBank accession, SRA BioProject, GEO dataset, PDB structure, Dryad dataset, Zenodo DOI, or Figshare record can be cited and reused, but it is not automatically a full scientific paper unless accompanied by a formally published article or data paper.

In the prior-publication sense

Usually, depositing required data in a recognized repository is not treated as prior publication of the manuscript. In fact, many journals require or strongly encourage it. PLOS requires authors to make all data necessary to replicate findings publicly available at publication, and PLOS Biology asks authors to deposit appropriate datasets in public repositories and provide accession numbers or dataset DOIs in the manuscript. (PLOS) Springer Nature strongly encourages public availability of supporting datasets and mandates sharing of some community-endorsed data types, with persistent identifiers such as DOIs or accession numbers cited in the article. (Springer Nature)

So, a useful distinction:

Object	Is it public?	Is it peer-reviewed article publication?	Can it count as prior publication?
GenBank accession	Yes	No	Usually no
Zenodo dataset DOI	Yes	No	Usually no
Data paper in a journal	Yes	Yes	Yes, for that data description
Full conference proceedings paper	Yes	Sometimes peer-reviewed	May count, depending on journal
Preprint	Yes	No formal journal peer review	Usually no, but policy-specific
Journal article	Yes	Usually yes	Yes

Database deposition is publishing in the sense of making public. It is not usually publishing in the sense of peer-reviewed article publication. The distinction is small, but it carries a cathedral on its back.

10. What about abstracting services? Are abstracts “published”?

An abstract in a conference booklet, conference supplement, or indexing database is public, but it is not the same as a full paper.

A conference abstract may establish that the work was presented. It may contain preliminary results. It may be citable. But many journals still allow later full publication if the full paper substantially expands the work, includes complete methods, full data, analysis, and discussion, and discloses the earlier abstract. Policies vary, especially when the conference output is a full peer-reviewed proceedings paper rather than a 250-word abstract.

Indexing an abstract does not create a new publication. It creates a searchable record of something already published or deposited. PubMed, Scopus, Web of Science, and DOAJ are discovery systems. They do not themselves transform weak work into strong work, nor do they convert an abstract into a full article. They help the scholarly ecosystem find, classify, and count things.

A paper can be:

published but not indexed,
indexed but not in the database you care about,
in PubMed but not MEDLINE,
in PMC full text but not necessarily MEDLINE-indexed,
assigned a DOI but not necessarily peer-reviewed.

These are separate badges. Do not confuse the medals on the uniform.

11. Who defines these things?

No single person. No single office. No single ancient council of owls, sadly 🦉.

Definitions are produced by overlapping authorities:

Authority	What it shapes
Journals	Article types, acceptance rules, prior-publication policy
Publishers	Ethical policies, preprint policy, data policy, copyright
ICMJE	Biomedical manuscript and publication recommendations
COPE	Publication ethics guidance for editors and publishers
NLM/MEDLINE	Indexing and journal-selection decisions for MEDLINE
Scopus/Web of Science/DOAJ	Database inclusion and curation criteria
Funders	Open-access, data-sharing, trial-registration mandates
Institutions	Thesis, repository, authorship, misconduct policies
Disciplinary communities	Norms around preprints, conference papers, datasets
Repositories/databases	Submission, accession, release, metadata standards

ICMJE recommendations are influential in biomedical journals, especially around authorship, overlapping publication, trial registration, manuscript preparation, and data sharing. (ICMJE) COPE provides publication-ethics guidance and emphasizes that editors must communicate clear policies, especially on issues such as prior publication and preprints. (Publication Ethics) NLM decides whether a journal merits inclusion in MEDLINE, using scientific and editorial quality considerations and external expert advice. (National Library of Medicine)

So when someone asks, “Is this already published?” the scientific answer is often:

“According to whose policy, for what purpose, and with how much overlap?”

That may sound slippery, but it is accurate.

12. A practical decision guide for researchers

Before submitting a paper, ask these questions:

Has the same full manuscript appeared elsewhere?
If yes, it is likely prior publication.
Was it posted as a preprint?
Usually acceptable, but disclose it and check the journal.
Was it part of a thesis?
Usually manageable, but check journal policy and disclose when asked.
Was it a conference abstract or full proceedings paper?
Abstracts are often acceptable. Full proceedings papers need careful checking.
Are the same data used in another paper?
Make sure the new paper asks a distinct question, uses appropriate analysis, and cites the related work.
Have data been deposited in a repository?
Usually good and often required. Cite the accession or DOI.
Was any text, figure, or table reused?
Cite, disclose, get permission if needed, and avoid copyright trouble.
Does the target journal have a specific policy?
The journal’s policy is the door lock. Do not arrive with the wrong key.

13. Why this matters

These distinctions are not bureaucratic hair-splitting. They protect the integrity of the scientific record.

Primary papers introduce evidence.
Reviews digest evidence.
Preprints accelerate visibility.
Databases preserve reusable objects.
Indexes make scholarship discoverable.
DOI systems connect the record.
Trial registries reduce selective reporting.
Prior-publication policies prevent duplication and distortion.

The scientific literature is not a bookshelf. It is an ecosystem. Some organisms produce oxygen. Some decompose old claims. Some pollinate new ideas. Some are invasive weeds with suspiciously glossy leaves. A good researcher learns taxonomy before wandering too deep.

Final thought: a paper is not just “published.” It is positioned.

The same research can exist in many forms: a conference poster, a preprint, a thesis chapter, a dataset, a protocol, a peer-reviewed paper, a review article, a database accession, a press release, and a policy brief.

They are not interchangeable.

A scientific paper becomes trustworthy not merely because it is public, indexed, or citable. It becomes trustworthy when its claims, evidence, methods, limits, and provenance are clear enough for the community to examine.

That is the real publication event: not ink on paper, not a PDF online, but evidence entering the public arena with its armor properly labeled.

Nagarjun's blog

Monday, July 6, 2026

What Is a Scientific Paper? A Field Guide to the Strange Little Creature Called “the Literature”