Updates and corrections

The following updates and corrections have been made to the CWTS Leiden Ranking.

June 21, 2017. The update of the 2017 edition of the Leiden Ranking that took place on June 19 has led to some minor errors in the indicators for a small number of universities. These errors have been corrected.

June 19, 2017. Update of the 2017 edition of the Leiden Ranking. The following corrections have been made:

  • Ulsan National Institute of Science and Technology in South Korea has been added to the ranking. This university incorrectly was not included. As a result of adding this university to the ranking, the total number of universities included in the ranking has increased from 902 to 903.
  • The indicators for Norwegian University of Science and Technology have been corrected. The merger with the university colleges of Sør-Trøndelag, Ålesund, and Gjøvik incorrectly had not been taken into account.
  • The indicators for Arizona State University in the US have been corrected. Publications from some campuses incorrectly had not been assigned to Arizona State University.
  • The indicators for Université Grenoble Alpes in France have been corrected. Some minor corrections have been made to the publication data for this university.

May 17, 2017. Release of the 2017 edition of the Leiden Ranking. The following changes have been made compared with the 2016 edition:

  • Indicators on collaboration with industry. These indicators have been reintroduced in the Leiden Ranking. They were available in earlier editions of the ranking, but they were not included in the 2016 edition.
  • Fractional counting method. In earlier editions of the Leiden Ranking, address-level fractional counting was used. In the 2017 edition, we have switched to author-level fractional counting. In the case of address-level fractional counting, each address in the address list of a publication has equal weight. In the case of author-level fractional counting, on the other hand, each author of a publication has equal weight. To illustrate the difference between the two fractional counting approaches, consider a publication has been co-authored by five researchers. Three researchers are affiliated with university X. The other two researchers are affiliated with university Y. The three researchers affiliated with university X belong to two different departments within the university. Both departments are listed as a separate address in the publication. The two researchers affiliated with university Y belong to the same department within the university. In the case of address-level fractional counting, used in earlier editions of the Leiden Ranking, the publication is assigned to universities X and Y with weights of 2 / 3 = 0.67 and 1 / 3 = 0.33, respectively. In the case of author-level fractional counting, used in the 2017 edition of the Leiden Ranking, the weights with which the publication is assigned to universities X and Y are, respectively, 3 / 5 = 0.6 and 2 / 5 = 0.4.
  • We have improved the procedure for assigning publications to Flemish universities. The identification of Flemish universities in Web of Science affiliations is particularly difficult because of the name conflation between Flemish and French speaking universities in Belgium. To cope with the complexity of ambiguous affiliations, the Leiden Ranking uses a special assignment procedure for Flemish universities. This procedure is different from the procedure that is used for other universities in the Leiden Ranking. In earlier editions of the Leiden Ranking, we assigned publications to Flemish universities as part of this procedure. In the 2017 edition, we have revised the procedure. Instead of publications, we now assign affiliations within publications to Flemish universities. This influences the weights with which publications are assigned to Flemish universities in the fractional counting approach. In the case of publications with multiple affiliations from the same Flemish university, the weight usually will be higher in the 2017 edition of the Leiden Ranking than in earlier editions. In the fractional counting approach, Flemish universities therefore obtain a substantially larger publication output in the Leiden Ranking 2017 than in earlier editions of the ranking. This provides a more realistic representation of the contribution these universities have made to the publications co-authored by them.

May 18, 2016. Release of the 2016 edition of the Leiden Ranking. The following changes have been made compared with the 2015 edition:

  • Selection of universities included in ranking. In the 2015 edition of the Leiden Ranking, the 750 universities worldwide with the largest Web of Science indexed publication output were included. In the 2016 edition, we use a different approach to select the universities that are included in the ranking. Rather than selecting a fixed number of universities, we include all universities worldwide whose publication output is above a fixed threshold. This threshold equals 1000 fractionally counted Web of Science core publications in the period 2011–2014. Using this threshold, 842 universities have been selected for inclusion in the 2016 edition.
  • Academic hospitals. The treatment of some academic hospitals has changed in the 2016 edition of the Leiden Ranking. In earlier editions, publications mentioning an affiliation with an academic hospital that is part of or controlled by a university were assigned to the university. In addition, publications from an academic health science center to which a university delegates its medical research and teaching duties were also assigned to the university. In the 2016 edition, publications from academic hospitals that do not satisfy the above criteria but that are nevertheless very tightly integrated with a university are also assigned to the university. Researchers that work at these hospitals but that are employed by the university often turn out not to mention their affiliation with the university. We assume that researchers do not mention their university affiliation because the hospital is perceived to be part of the medical faculty of the university.
  • Inter-institutional collaboration indicators. In the 2015 edition of the Leiden Ranking, a collaboration between a university and an affiliated organization that is considered to be part of the university was regarded as an inter-institutional collaboration. For instance, a collaboration between Leiden University and Leiden University Medical Center was seen as an inter-institutional collaboration. In the 2016 edition, these collaborations are no longer regarded as inter-institutional collaborations.
  • International collaboration indicators. In the 2015 edition of the Leiden Ranking, England, Northern Ireland, Scotland, and Wales were regarded as separate countries in the calculation of the international collaboration indicators. In the 2016 edition, the United Kingdom is regarded as a single country in the calculation of the international collaboration indicators. Hence, a collaboration between for instance an English and a Scottish university is no longer seen as an international collaboration.
  • Indicators on collaboration with industry. These indicators are no longer available in the Leiden Ranking. Indicators on universities’ collaboration with industry will be published separately in 2017.
  • Website. The Leiden Ranking website has been significantly revised. In addition to a traditional list-based presentation, where universities are presented in a list ranked based on a selected indicator, the revised website also offers a chart-based and a map-based presentation. Furthermore, in the list-based presentation on the revised website, size-dependent and size-independent indicators are always presented together and the P indicator has replaced the PP(top 10%) indicator as the default indicator for ranking universities.

Share this page


Contact