Scientific publications are not limited to journal articles. They include various formats, and each of them plays its own role in academia. Understanding their details is important when writing your own research, and also when assessing scholarly outputs, searching for information, or preparing a publication.
Peer-reviewed and non-peer-reviewed publications: what is considered as "scientific"?
Traditionally, the key characteristic of a scientific paper is its peer-reviewed status: when other specialists in the field provide their expert evaluation of the work. Peer-reviewed publications are most commonly taken into account in accreditation processes, grant applications, and bibliometric analyses. However, other types of publications are also essential, particularly within the context of open science trends and interdisciplinary collaboration.
Thus, publications can be classified into two main categories:
- Peer-reviewed: scientific articles and reviews, conference proceedings, books and book chapters, papers in collected volumes, and monographs.
- Non-peer-reviewed: preprints, popular science articles, technical reports, news and blogs, abstracts, dissertations, and textbooks.
The publication types differ in terms of editorial policies, peer review procedures, and adherence to formal academic standards.
For example, a preprint provides rapid dissemination of results, however, it lacks formal peer-reviewed status. On the other hand, a dissertation represents a comprehensive scholarly study, however, it is not always accessible to a wide audience.
What do the databases show?
Remember that the extent and the quality of indexing significantly varies depending on the publication types in scientific citation databases. Web of Science and Scopus primarily cover peer-reviewed journal articles and conference proceedings, however, poorly index monographs, books, book chapters, and datasets. Google Scholar offers a broader coverage, including preprints and other open-access materials, however, applies less rigorous criteria for source selection.
For example, according to Scopus data (as of 2025), the vast majority of records in the database are journal articles, conference abstracts, and reviews. Articles account for over 50 million entries, conference papers exceed 11.6 million, and reviews number around 4.5 million. In contrast, other forms of scholarly outputs are represented much more modestly:
- Book chapters — 3,970,905
- Books — 447,361
- Datasets — 19,934
- Reports — 2,475
This imbalance leads to a systematic underrepresentation of the disciplines where alternative publication formats prevail. For instance, humanities and social sciences are poorly shown in Scopus and Web of Science, as research outputs there are often published as monographs, book chapters, or similar formats. On the other hand, researchers’ significant contributions, including data creation, software development, patents, and books, remain invisible to conventional bibliometrics. Therefore, when evaluating scholarly activity, it is essential to consider a broader spectrum of academic impact rather than only articles.
In fact, the same scientific work may exist in multiple formats: as a preprint, a conference abstract, and only later as a journal article. Citation databases handle the indexing of versions differently: Google Scholar typically groups them together, while Scopus and Web of Science often treat them as independent materials. The Russian Science Citation Index (RSCI) has successfully addressed the issue of linking Russian- and English-language versions of the same paper, avoiding citation duplication.
Main types of peer-reviewed publications
1. Research Article (Article)
This is the most common type of scientific publication. It provides a complete description of a study, including the research issue, methodology, results, and conclusions. Research articles present a discovery, a method, a technology, a theoretical model, or another form of new knowledge. They are considered as primary sources in scholarly communication.
2. Review Article (Review)
This is an analytical summary of research within a specific field. It helps scholars quickly grasp the state of the art, identify knowledge gaps, and outline the directions for future studies. These publications are often among the most highly cited works.
3. Proceedings Paper / Conference Paper
Extended abstracts or full short papers, published in conference proceedings or special journal issues. They may offer less depth than full journal articles, however, they play a crucial role in the rapid exchange of ideas and dissemination of preliminary findings, especially in engineering, computer science, and other fast-moving fields.
4. Monograph and Edited Volume
They represent a detailed scholarly work on a single subject, written by one or several authors, representing an in-depth and comprehensive study.
An edited volume (or collected volume) is a collaborative publication in which authors contribute their individual chapters on related topics, coordinated by one or more editors who ensure thematic coherence and academic quality.
5. Book or Book Chapter
These are used to systematize the knowledge on some complex subjects, including fundamental research, comprehensive reviews, and educational content. Peer-reviewing for books is less standardized than for journals, however, the academic quality remains high. Book chapters contribute focused analyses within broader thematic contexts.
New Format: Datasets – When Data Becomes Publication
With the advancement of digital technologies and the growing popularity of the open science movement, a new format of scholarly publication has emerged, that of Dataset / Data paper. Today, fully annotated data collections are increasingly recognized as independent scientific outputs that undergo rigorous peer review.
Datasets have gained a status comparable to that of traditional publications. The primary value of a dataset lies in its transparent description and structured arrangement rather than in the analysis of the data. Being available, data papers enhance transparency, support validation of results, and foster further discoveries across disciplines.
Transforming data into an independent scientific publication makes it a full-fledged component of scholarly discourse. This is important for several reasons:
- It makes data discoverable and citable: assigning DOI and other persistent identifiers can trace research outputs and reference in academic work.
- It supports the FAIR principles: data are Findable, Accessible, Interoperable, and Reusable.
- It promotes transparency and reproducibility: other researchers can verify results and replicate studies.
- It provides recognition to scholars who arrange data: the efforts involved in collecting, curating, and documenting.
Where are Datasets published?
This type of publication is accepted in specialized journals. Here are some examples:
- Scientific Data (Nature) is a leading journal focused on the description of scientifically significant datasets.
- Data in Brief (Elsevier) publishes concise data articles that describe datasets related to primary research.
- GigaScience (BMC) integrates research articles with large-scale, high-volume datasets, emphasizing data reuse and reproducibility.
- IEEE Data Descriptions focuses on technical and engineering data, providing peer-reviewed descriptions of data collections.
- Geoscience Data Journal, BMC Research Notes are subject-specific platforms supporting data publication in various disciplines.
Authors also publish their data in Figshare, Dryad, Zenodo, and other trustworthy, open-access repositories, ensuring long-term availability, preservation, and citability supported identifiers.
Why is this important to know?
The growing diversity of publication formats reflects the evolution of science: from closed, article-centric communication to open, reproducible, and collaborative knowledge production. With the informed choices on the publication types researchers can share results more effectively, gain proper academic recognition, and make meaningful contributions to the advancement of their fields.
