The researchers looked at the grades achieved in nine English-language bachelors, spread across all the faculties at Maastricht, between 2015 and 2021. The results were published last month in the scientific journal Higher Education.
It is not a direct response to the most recent government’s discussions about internationalisation in higher education, said Patrick Bijsmans, associate professor at the Faculty of Arts and Social Sciences (FASoS), who carried out the research together with researchers from the Research Centre for Education and the Labour Market (ROA). “We started researching this in 2019, before the political debate really took off. Based mainly on the thought that, as a university, we assume that an international classroom works well, but does it really? There has been very little research on the subject. At the same time, you could see more and more assumptions arise in politics, such as that the level of education and the performance of Dutch students was going down as a result of more international students. We could counter that anecdotally, but had no data to back it up.”
Random groups
Maastricht is the ideal place to test the idea that students score higher “if they come into contact with more different perspectives, instead of staying in their national bubble”, said Bijsmans. Not just because the number of internationals is high here, but also because of the smaller set-up of problem-based learning (PBL). “It means students really have to interact with each other,” said professor Carla Haelermans of the ROA. “Furthermore, on most courses, they are split into groups randomly. They can’t choose to be in a tutorial group with only friends from their own country.”
Thanks to this random distribution, the number of different nationalities in the groups for one particular course can vary greatly. “On average, we saw that groups with, say, three to six different nationalities achieved higher grades than those in a group with only one or two different nationalities,” said Haelermans. “Although there was an upper limit, too. If there are, for example, ten different nationalities in a group, the effect disappeared again.” That was the case in the Maastricht Science Programme, where the groups are not assigned completely random, but the diversity is deliberately as broad as possible. “When there are slightly fewer nationalities, the chance that you are in a group with a compatriot is bigger, which can also provide some support. For example, if you can’t quite find the right word or run into cultural differences.”
Common language
Nevertheless, the consistent use of a common language is important. In groups where teachers stimulate the use of English, even during breaks, the effect was stronger. “It’s about balance,” said Bijsmans. “On the one hand, there is an opportunity to check whether you understood something correctly with a compatriot in your own language, but at the same time, keeping the conversation going for everyone.”
Not every student benefits though. There is little difference for the average German student when the number of nationalities increases. “Although there could be other reasons for that,” said Haelermans. “For example, at the School of Business and Economics, where there are many Germans, they use selection. So the average German student there is highly motivated and already achieves higher grades. That leads to a ‘ceiling’ effect: it is hard to further improve a grade that is already high.” According to Bijsmans, there is little point only focusing on specific nationalities within PBL. “In the end, it’s about improving the performance of the whole group.”
Apples and oranges
Opponents of internationalisation might object to the fact that the results of Dutch students in ‘international’ groups were not compared to those of students surrounded only by other Dutch students. Why weren’t they? “At Maastricht, groups consisting solely of Dutch students only happens in Dutch-language courses,” said Haelermans. “Their results cannot be compared to those of groups for the same course with multiple nationalities, as they simply don’t exist. And comparing the difference in results between two different programmes is like comparing apples to oranges.”
It is also not the case that UM – who funded the research, including through a Diversity and Inclusivity grant – were looking for a favourable outcome, said Bijsmans. “The university is interested in what works and what doesn’t. There is no earlier research that has been able to examine this much data.” There’s no inherent conflict of interest, added Haelermans. “A researcher in Groningen examining the data would reach the same conclusions. We were very thorough.”
Finally, is there anything the university could do to improve? According to the two, the way groups are assigned could be changed. “At random is good, but try to prevent creating groups with too few or too many different nationalities.” More in-depth research currently in the pipeline, in which the researchers observed tutorials and interviewed teachers and students afterwards, has “highlighted other areas of attention”, said Bijsmans. “For example, you could train teachers to take better advantage of the diversity of a group, so that students can give voice to their own perspectives more. That currently doesn’t always go as well as it could.”