Universities in large cities create less highly cited research than institutions elsewhere, possibly because high rents and living costs put off the best staff and students.

This is one of several surprising findings of an analysis that overturns several commonly held ideas about why some universities excel at research.

One popular notion is that in major cities with several universities ¨C places such as London, Hong Kong, Boston, New York or Paris ¨C researchers find it easy to mingle with other experts in their field through formal collaborations and at social gatherings, sparking new ideas.

This networking effect likely does benefit universities located in metropolises, said Gaston Heimeriks, an assistant professor in innovation studies at Utrecht University and co-author of the analysis, but it could be outweighed by the expense of living in these conurbations, where costs make it difficult ¡°to attract the best staff and students¡±.

É«ºÐÖ±²¥

The analysis, which looked at the research record of 750 universities between 2010 and 2013, found that universities in bigger cities did not produce more highly cited papers than others ¨C if anything, the number of such papers dipped very slightly as the population increased.

In dense urban areas, universities can face ¡°more planning constraints and higher costs¡±, while ¡°students and staff may be reluctant to work in big cities because the living costs are higher, whereas wages are often determined nationally¡±, the analysis suggests.

É«ºÐÖ±²¥

This tallies with conducted by Professor Heimeriks that found that in the US, the very biggest cities produce fewer scientific papers than might be expected. ¡°Very large cities are unattractive locations for scientific research,¡± that work concluded, perhaps because of high rental costs.

In recent years, universities in expensive cities such as Cambridge, London, Hong Kong and San Francisco have grappled with how to attract academics to their pricey cities, and many have turned to providing subsidised housing.

Search our database for the latest global university jobs?

The analysis also finds that technical universities, which normally concentrate on the vocational end of research, actually create about 7 per cent more highly cited papers than traditional universities, when other variables are held constant.

This surprised Professor Heimeriks, who suggested that it could be because ¡°a lot of new and exciting scientific developments are taking place in fields that are traditionally done by technical universities¡±, such as information technology, nanotechnology and energy.

É«ºÐÖ±²¥

The analysis ¨C ¡°¡±, published in the Journal of Informetrics ¨C also finds that hardly any countries boast universities that excel at producing research that is simultaneously highly cited, has international co-authorship, and is produced in collaboration with industry.

At best, most countries manage to do well on two out of three of these metrics (the UK, for example, creates research that is highly cited and international but lacks industry collaboration).

¡°Given limited resources and incentive structures, there¡¯s a trade-off between these activities,¡± said Professor Heimeriks. Governments want their universities to be ¡°five-legged sheep¡± that excel at everything, he said, but the analysis suggests that this might not be possible.

The one exception is Switzerland, which performs very highly on all three measures. ¡°My guess is that Switzerland has a very unique university structure,¡± he explained. ¡°They have concentrated all their research into a few universities,¡± and the country has historically very strong links to surrounding countries and a deep industrial base, he said.

É«ºÐÖ±²¥

david.matthews@timeshighereducation.com