Bibliometrics

Bibliometric analyses as well as indicators are based on data, such as literature references and citations, which can be determined via various databases. It should be noted that the significance and meaningfulness of a bibliometric analysis strongly depends on the quality, coverage and orientation of the database. For example, the H-index of individual scientists differs depending on which data source is used as a basis.

Web of Science (WoS) is a multidisciplinary literature and citation database focusing on the natural sciences. It is operated by the company Clarivate Analytics. The Web of Science lists scientific publications as well as the related citations. In addition, various bibliometric metrics are available. The University Library of the TU Berlin has licensed Web of Science (until the end of 2023), access is free for university members.

Scopus is a multidisciplinary literature and citation database operated by the Elsevier company. The subject areas covered are strongly focused on the natural sciences. In addition to literature references and citations, various bibliometric metrics are also available via Scopus The University Library of the TU Berlin has licensed Scopus, access is free for university members.

Dimensions is a multidisciplinary literature and citation database operated by Digital Science. In addition to scientific publications and their citations, this database also lists research data, grants, funding organizations, and patents, among other things. In addition to licensable access, Dimensions is also available via free access (with limited features). The University Library of the TU Berlin has not licensed Dimensions

The free database OpenAlex lists and links scientific publications, institutions, authors, concepts and citations. OpenAlex follows open source principles and makes its entire index of data openly accessible via a web interface, an API and a database snapshot.

Google Scholar is a publicly accessible search engine for academic literature, with a very large database showing citation details for the results it shows. The data sources, however, are not transparent and results are not reliably reproducible.

The basis for bibliometric indicators is usually the number of publications and their citations.

Author indicators are designed to evaluate and compare the scientific output of an individual researcher. TThey do not replace a qualitative evaluation of the scientific achievement of a  researcher.

Very basic bibliometric indicators are the number of publications or the number of citations of a single researcher.

The citation rate indicates the average number of citations per publication of a person. The citation rate is the ratio of the number of publications (P) to the number of citations (C): Citation rate CCP = C/P.

The H-index, also known as the Hirsch index, was developed to evaluate and compare the scientific output of individual scientists. It is calculated from the ratio between the number of publications of a person and the frequency of citations of the articles. It is defined as the number of articles h of a person that have been cited at least h times. For more information on the H-index, see the post „Der H-Index – ein besseres Maß?“ in our blog.

The Field-Weighted Citation Impact (FWCI) is calculated by the Elsevier for an author and is distributed commercially through the Scopus database. It indicates how often the publications of an entity, compared to an expected value of the same discipline, have been cited. The expected value refers to the average number of citations in the last three years for publications of the same age, document type and subject area.

Journal indicators are designed to measure the influence of individual journals. They are not indicators for evaluating the output of an academic. They also provide no indication of the quality of an individual article in a journal. An article published in a journal with a high journal indicator is not necessarily qualitatively better than an article published in a journal with a lower indicator.

The best known and oldest indicator is the Journal Impact Factor (JIF). It indicates how often an article from a journal has been cited on average in other publications. It is calculated from the number of citations of a journal's articles in the year under consideration in relation to the number of these articles in the previous two years.

The Journal Impact Factors are calculated once a year by Clarivate and published in the Journal Citation Reports (JCR) as a commercial product. The journals recorded in the Web of Science database (from the "Core Collection") and the citation frequencies recorded therein serve as the basis for calculation. This means that only journals that are indexed in Web of Science receive a Journal Impact Factor. For more information on the JIF, see the post „Der Journal Impact Factor – ein geeignetes Maß?“ on our blog.

The CiteScore, similar to the JIF, indicates how often a journal article was cited on average. The calculation of the CiteScore is based on the number of citations of documents (articles, reviews, conference papers, book chapters, and data artiecl) of a journal over four years, divided by the number of the same document types published in the same four years. Journals and citations indexed in Scopus serve as the data basis. Access to the CiteScore (as well as other indicators, including SJR and SNIP) is free.

Eigenfactor determines the scientific influence of journals using mutual citations of articles (network analysis). Based on the available citation data (data basis: Web of Science), two metrics are calculated using the Page Rank algorithm: the Eigenfactor Score (ES) and the Article Influence Score (AIS). Both are freely available.

To reflect new forms of science communication and the growing number of electronic publications, article-related, alternative metrics are being developed on the basis of use data (e.g. views, downloads, bookmarks) and discussions of academic publications in social networks (e.g. Twitter, Facebook, blogs).  These alternative metrics (altmetrics) are provided by different providers, such as Altmetric.com (Digital Science) or Plum Analytics (Elsevier).

Author profiles assist with the clear identification of authors. They ensure all publications are correctly attributed to a researcher. This promotes a comprehensive evaluation and visibility of the researcher’s publication output.

The University Library recommends all researchers register with ORCID and use their individual ORCID in their publications. More about ORCID

Open access publications are available free of charge and publicly online, in other words without any financial, legal, or technical barriers. This allows open access contributions to be found and cited more often than content that is not publicly accessible online. More about open access

Science is a competitive field and including an author’s affiliation to an institution in publications is of key importance. Universities and their researchers are frequently ranked in national and international comparisons according to their publication output, with third-party funding often awarded on the basis of these rankings. Clear and, above all, complete details are also important for other studies evaluating data on affiliation, e.g. analyses of inter-institutional cooperation.

In light of this, TU Berlin approved a set of guidelines for standardizing statements of affiliation in German and English-language publications in October 2019. These guidelines apply to all members of TU Berlin. They determine the use of a standardized name for the University (Technische Universität Berlin) as well as a standardized abbreviation (TU Berlin) and establish the procedures to be followed in the event of affiliation to more than one institution.

The guidelines ensure that all publications produced by researchers at TU Berlin are clearly, correctly, and fully assigned to Technische Universität Berlin. As such, they serve to increase the visibility of TU Berlin's research strengths and help ensure that the University’s research attracts funding. They also have a positive impact on national and international rankings.

You can read more about TU Berlin’s affiliation guidelines in our blog article: "TU Berlin beschließt Affiliationsrichtlinie” (German only)

Bibliometric analyses are useful for quantitatively measuring, evaluating, and comparing academic publications. They are increasingly used in science management, for example as an instrument to identify research trends or awarding merit-based funds.

Bibliometric analyses are solely an instrument of research management. They are never a replacement for a qualitative evaluation of research content. This is possible through peer review, the review of research by independent experts, for example.

The most important bibliometric data sources are the multidisciplinary literature and citation databases Web of Science and Scopus. It is important to note that the validity and meaningfulness of a bibliometric evaluation strongly depend on the quality and focus of the database. A researcher’s h-index can vary depending on which data source was used as a basis.

Individual scores should never be interpreted in isolation. Rather, the entirety of all indicators is to be considered and reasonably evaluated. For example, the number of publications (output) says nothing about the quality of the publications. Someone who publishes frequently is not necessarily publishing at a better qualitative level, just as someone who publishes less often is not necessarily publishing at a lesser qualitative level.

As the publication activities of academics strongly differ in individual disciplines, bibliometric indicators are not suitable for cross-disciplinary comparison.