There are clear benefits to be had from international collaboration in research. Researchers can access skills and experience, data and equipment not available at home, which can open up opportunities for future development.

But does international collaboration translate into a tangible boost in the reach, recognition and academic impact of ones research? One (crude) way of measuring this is via citations. To explore this, I used the SciVal database[1] to investigate the possible correlation between international collaboration[2] and citations.

To measure citations, I used the Field Weighted Citation Impact (FWCI) metric. FWCI is a measure based on the number of citations that publications receive.  It’s normalised to account for subject area differences, when a publication was published,and the type of the publication. This means that it’s fairer than looking at raw citation counts, which can be biased towards certain subject areas and publications published a longer time ago which have had more time to accrue citations.

Each dot on the graph below represents a UK university. Portsmouth is red, other universities in the Alliance Group are yellow, universities in the Russell Group group are green, and the blue dots are all other UK universities.

There’s a significant positive correlation between international collaboration and FWCI, as shown on the graph above and the detailed results. This indicates that more international collaboration could lead to a higher citation rate in the long-term. This finding reflects previous work, which also reported a similar correlation[3][4].

To explore this in more depth, I focused in on a group of comparators. This includes all Alliance group universities, plus a couple from outside the group to put things into context. The graph below shows a clear pattern. For all universities except one, FWCI is higher when publications have international collaboration.

This pattern also holds when looking at articles submitted or proposed for the Research Excellence Framework (REF). As the graph below shows, publications submitted to the REF by Portsmouth in 2014 which had international collaboration also had a higher average FWCI.  This pattern holds for publications proposed for the University’s Mock REF exercise in 2018.

Previous work that analysed the REF 2014 results for the whole country also found that international collaboration and FWCI were predictors of higher REF scores[5].

While these University-wide results are interesting, they don’t answer the question of whether the pattern occurs across all subject areas.

Splitting the data into 11 broad subject areas, including Arts and Humanities, Business and Economics, Clinical and Health, Engineering and Technology, Life Sciences, Physical Sciences and Social Sciences, there is a significant positive correlation between percentage of international collaboration and FWCI in almost every area. (Results in detail).

The only exception is in the subject area of Law, where there was no significant correlation between the two measures. This may be because of the fundamental role of the legal jurisdictions that apply in different countries, meaning that publications can still secure high levels of citations by focusing on issues in one country’s legal system.

What’s more interesting, however, is that while all subjects (except law) have a positive correlation between international collaboration and FWCI, the ‘range’ in which this correlation operates varies. On average STEM subjects have a higher percentage of international collaboration compared to social sciences and the arts, but the positive correlation is still present in almost all subjects. (Graphs of the results). This suggests that it would not be realistic to expect all subject areas to reach the same percentage of international collaboration, but almost all subject areas may benefit from international collaboration to some degree.

Finally, I should acknowledge that FWCI is only one metric.  While this metric is important, it should be considered within the broader context and be aware that some disciplines, especially the arts and humanities, have other factors to consider. So the purpose of this short blog post is to show the overall trend in the data, as a starting point for further discussions.

Please contact Emily Davey (Research Outputs Manager, University Library) or Katie Osgood (Research and Innovation Services) if you have any further questions or thoughts.

 

References

[1] Data source: SciVal is a comprehensive database of publications, including journal and conference articles.
[2] ‘International collaboration’ refers to publications (e.g. journal articles) that have at least one co-author affiliated with a non-UK organisation (e.g. a non-UK university, business, government department etc).
[3] Influence of international co-authorship on the research citation impact of young universities
[4] International comparative performance of the UK research base 2016
[5] The Metric Tide Correlation: Analysis of REF2014 scores and metrics