Selection criteria and presentation
Karolinska Institutet is Sweden’s largest forum for academic research in the field of life science. Our researchers produce a wealth of scientific articles on everything from global public health issues to healthcare development and basic research into the very smallest components of life. A very large proportion of these publications are built on collaboration with the healthcare sector and other research stakeholders, both in Sweden and internationally.
During May 2019, we are upgrading the web system (CMS) for ki.se. The aim is to make the transition as smooth as possible. However, disturbances may occur in the information feed from the publication database of the University Library.
A list of top publications has been in existence in some form or other at KI for more than a decade, presenting current scientific articles to which our researchers have contributed, including both original results and review articles. For many years, the list was updated manually and the criteria for inclusion were that at least one of the authors was affiliated with KI, and that the article had appeared in a journal with impact factor (JIF) of 15 or higher.
From summer 2018, the top publications list will be updated by automatically retrieving information from the Medline and Web of Science databases at roughly two-week intervals (with a somewhat longer updating period over the New Year). At the same time as this relaunch, the criteria for inclusion on the list have been broadened somewhat.
These new inclusion criteria have been developed in consultation with KI’s scientific management, the University Library bibliometrics group and the Communications and Public Relations Office. Please note that selection criteria may be different for press releases and popular science news published on KI’s websites.
Criteria for inclusion on the top publications list
In order to be included on the top publications list, an article must have at least one KI researcher listed among the authors, and be published in a scientific journal that meets at least one of the following criteria:
- The Journal’s JIF is within an interval corresponding to that for the top 5% of all articles in Medline. The top publication list threshold for 2019 is a JIF of 9.66.
- The journal’s Journal Citation Factor (JCf), i.e. its field-normalised citation score, is within an interval corresponding to that for the top 5% of all articles in Medline. The top publication list threshold for 2019 is a JCf of 2.62.
The limitation to Medline means that these threshold values are based on the university’s main fields, although articles in other fields may also be included in the top publication list.
Because the JIF and JCf for the various journals are measured at different times and at varying intervals, for the sake of clarity we have decided to establish a list of top-ranked journals annually, at some point during September-October. This list of publications will then form the basis for the automated flow of scientific articles to the top publications list over the following year.
When an article’s reference flows into the top publications list, the level of affiliation to KI will also be graded; as low, medium or high respectively. A low affiliation to KI requires only that one of the listed authors has stated some form of connection to KI. A medium level of affiliation means that at least five, or alternatively at least 10%, of the listed authors are affiliated to KI. A high level of affiliation requires that the article’s corresponding, first or last author is active at KI.
It is not possible to manually add or remove articles from the automatic flow into the top publications list. Articles on the chart are linked to PubMed (Medline) via KI’s University Library and not directly to the journal’s own website as previously.
Other limitations compared to the previous manual list are that it is not possible to include the various journals’ own publication vignettes, for example Letter, Original Research or Analysis, and sometimes it may not be possible to include the date of publication. In some cases, article can be posted in the list several month after the actual publication date, since registration in article databases is connected to journal editorial routines. For the same reason, it is often the date of the print issue that is specified in the list, instead of the firs online publication date.
A way for researchers to speed up the top-list publication process is to verify their own articles in the University Library’s bibliometrics database.