The International Research Network (IRN) was included for the first time in the 2016/17 edition of the QS University Rankings: Latin America, and has now been expanded to all the other regional rankings.
Since 2022, we have included IRN Index in the ranking of broad subject areas (Arts & Humanities, Engineering & Technology, Life Sciences & Medicine, Natural Sciences, and Social Sciences & Management).
Also, IRN Index methodology is the ground of "Knowledge Exchange: Progress / Dissemination" metric of QS World University Rankings in Sustainability.
The Margalef Index, widely used in the environmental sciences, has been adapted to estimate the richness of the selected international research partners for a given institution. IRN Index reflects the ability of institutions to diversify the geography of their international research network by establishing sustainable research partnerships with other higher education institutions. It also reflects the efficiency of this as we look at the diversity of partner locations against the efforts needed to achieve such a diversity. Specifically, the QS International Research Network (IRN) Index is calculated with the following formula:
IRN Index = L / ln(P),
where P is the distinct count of international partners (higher education institutions) and L is the distinct count international locations represented by them.
In QS World University Rankings, QS University Rankings by Region, as well as QS Rankings by Subject, this metric considers only sustained partnerships, i.e. those which result in 3 or more joint papers with non-zero citations (excluding self-citations) indexed by Scopus in a five-year period. We produce IRN Index for only those subjects with a decent volume of papers (those ones with a paper threshold more than 50 over the last 5 years).
As in other cases when citations and papers are processed, IRN Index is normalized by our five faculty areas: it is calculated per each of them and then averaged to ensure that the influence of research publishing cultures is equalized across the different faculty areas. Also, only relevant paper types are considered and our usual affiliation cap is applied.
Note on the QS World University Ranking 2023/2024 editions of this indicator:
When we transform an institution’s scores into Z-scores, we lock the mean and standard deviations that we use in our sample. We do this so that institutions do not receive score increases simply because we have added new institutions to a ranking year on year. When we add new institutions, their performance across indicators is typically rather low. Adding new institutions may move them closer or further away from the mean performance, and thus adjusting their rank position based on no underlying shifts in performance. For International Research Network, we set the sample size to the top 700 – as we do for Citations per Faculty. This also helps ensure a more equal distribution of scores across indicators- making it easier to compare, e.g., Citations per Faculty with International Research Network.
Last year we published this indicator unweighted, the intent being to give a rough estimation of how an institution would fare. We did not, however, apply our standard normalization procedures for that indicator. This year we have reconciled the normalization in line with our other indicators. Institutions may compare rank positions rather than scores for this indicator and for this edition. Further details on this approach here: https://support.qs.com/hc/en-gb/articles/4402503754130-Z-Score-