In recent years, “academic reorganisation” policies have emerged in various countries. In countries such as Japan, France, and Germany, these initiatives encourage the grouping together of universities, and differentiating them by assigning ‘hierarchical’ roles (global, national, local, or research or only teaching). The main focus of these policies is to allocate resources to universities based on diagnostics and evaluations. A number of these assessments are based on ideas about the supposed spatial distribution of higher education and research.

These widespread beliefs about the nature of research include: the idea of ​​an ineluctable increase in the hierarchy among cities, the existence of a critical mass effect imposed by increasing globalization, and the idea that research is competitive and will shape the future of national economies.

I participated in a bibliometric study on the distribution of scholarly publications among the cities of the world and the collaborative networks within which those studies were done, along with a group of researchers from the LabEx project “Structuring Social Worlds” (Denis Eckert, Marion Maisonobe, Béatrice Milard, Laurent Jégou, Josselin Tallec). Working with the CIRST in Montreal (in particular Yves Gingras and Vincent Larrivière) and the Observatoire des Sciences et des Techniques (OST), we geocoded all the publication data on the Web of Science, run by Thomson Reuters. We aggregated data by city. Based on this data, we then tested a number of very common beliefs about the geography of research.

Concentration

Compared to the world’s population, research activity as measured by the number of publications identified in the bibliometric databases is undeniably much more geographically concentrated. The countries historically involved in research – Europe and North America, excluding Mexico – account for a considerable share of publications (around 60%), while their share of the world’s population is much lower (around 15%).

Within each country, the gap between the distribution of publications and the population also varies considerably. Looking only at the 11 countries that produce the greatest number of articles according to the Web of Science, there are surprising differences: three countries have a much accentuated hierarchy with more than half of their scientific output in only two cities (Japan, Australia, and Russia). In other countries, it takes as many as 8 cities (India), 10 (Germany) or 19 (United States) to arrive at this threshold of 50%. Other countries, such as France, China, and the United Kingdom, are in the middle of these two extremes.

Deconcentration

In many contemporary discourses on science and the dynamics of academic activity, the idea that research is increasingly concentrated in a small number of cities is often widely accepted. A simple way of assessing the degree of concentration of research is to calculate the proportion of publications around the world that can be attributed to the world’s largest research cities.

Over a relatively short period (less than a decade), the contribution of major world metropolises to academic output has decreased. The top 10 cities account for just over 15% of the total research output, losing more than two percentage points compared to the beginning of the decade. This is a generalised tendency, whether we consider the thirty, fifty, one hundred largest cities in the world, or further down in the classification. This means that research activity is in the process of becoming more diffuse. This process is not recent, but slow and continuous. We have observed this trend since the late 1980s, when the 10 largest cities still accounted for 21% of the total. They have therefore lost five points in twenty years – proof that the theory of the concentration of research in large cities does not stand up to analysis.

Critical mass does not exist

It is nearly impossible to count the research papers and books on research practices that are based on what can be called “the theory of critical mass.” This metaphor means to say that a sufficient density of researchers is required in an institution, a city, or a region for the quality of the research to be good. Researchers are supposed to need many colleagues close by in order to exchange ideas and be intellectually stimulated.

Some attempts have been made to establish a link between the number of researchers in the same city or region and the average number of articles published per researcher. Yet these studies have not been able to find such a connection, as the total number of publications in a city or region depends mainly on the overall number of researchers, and that number results from changes in higher education and policies conducted at national or local level. The concentration of researchers has no effect on the number of their scientific publications – critical mass does not exist here. Other ongoing studies on the number of times an article is cited show a slight increase in visibility for large cities, but they also show that even that difference is decreasing.

International and national collaboration

Another widespread belief in the discourse on the spatial organization of research is the tendency towards the internationalization of collaborations and the regression of national contexts. Our data on cities allow us to study research networks by identifying the co-authors of research articles.

National collaborations have increased at the same rate as international ones, and both these changes are occurring to the detriment of publications authored by teams from a single city, and more specifically, publications from the same institution. Contrary to what one might think, over the last 10 years there has not been a global trend of researchers particularly trying to collaborate with international colleagues. The growth in international collaborations is part of an overall increase in collaborations between researchers from different cities, regardless of whether they are from the same country.

Diversification and de-concentration

These studies have shown that the beliefs guiding many ‘territorial’ research policies are rarely true: it has been proven that research is highly concentrated, but the tendency is more toward greater diversification and deconcentrating than strengthening the largest research cities. The concentration of researchers in a given city has no direct effect on the amount researchers publish. Moreover, national research is not disappearing, but rather combining with the increase in international collaboration in a global context in which studies by one single research team or institution are regressing. Here, spatial concentration has no particular effect on how much researchers publish. Finally, national research is not fading away; but is rather combined with the overall increase in international collaborations in a global context in which publications by a single research team are declining.

These beliefs have developed in a context in which the geography of research is still an emerging discipline. This quick overview has allowed us to show the importance of these kinds of studies.

Crédits image à la Une : CC Alpha Stock Images Nick Youngson et crédits image d’entrée : CC Pixabay Clker-Free-Vector-Images

---

---