Date Published: January 23, 2017
Publisher: Public Library of Science
Author(s): Lorenzo Cassi, Raphaël Champeimont, Wilfriedo Mescheba, Élisabeth de Turckheim, Pablo Dorta-González.
This paper shows how the Rao-Stirling diversity index may be extensively used for positioning and comparing institutions interdisciplinary practices. Two decompositions of this index make it possible to explore different components of the diversity of the cited references in a corpus of publications. The paper aims at demonstrating how these bibliometric tools can be used for comparing institutions in a research field by highlighting collaboration orientations and institutions strategies. To make the method available and easy to use for indicator users, this paper first recalls a previous result on the decomposition of the Rao-Stirling index into multidisciplinarity and interdisciplinarity components, then proposes a new decomposition to further explore the profile of research collaborations and finally presents an application to Neuroscience research in French universities.
Interdisciplinarity refers to complex processes of knowledge production that involves cognitive dynamics as well as a social construction [1–3]. It also covers a variety of ways to confront and bridge disciplinary approaches. Among the various definitions and terms that have been proposed in a now long standing debate, the distinction between multidisciplinarity, as a juxtaposition of disciplinary components, and interdisciplinarity which integrates knowledge, theories or methods from different disciplines is widely shared . However, there is not a single way to characterize interdisciplinarity as one could consider the degree of disciplinary integration, the interdisciplinary practices or the rationales of interdisciplinarity [5, 6]. Understanding interdisciplinarity therefore deserves combined methods involving an analysis of the integration process as well as quantitative measurements based on outcomes [1, 3]. Among quantitative methods, the bibliometric literature proposes different indicators, based on different data, different levels of aggregation and different visualisations adapted to different needs of indicators users.
Neuroscience—also referred to as neurosciences—is the study of the structure, functions, development, abnormalities of the nervous system and its impact on behaviour and cognitive functions both in the normal functioning and in the case of neurological, psychiatric and neurodevelopmental disorders. Neuroscience is currently a multidisciplinary field involving biomedical sciences—such as clinical neurology, psychiatry, cognitive and behavioural science -, fundamental biology—such as genetics and molecular biology -, as well as other disciplines as psychology, linguistics and philosophy, together with engineering, chemistry, physics, mathematics and computer science. Due to the large range of disciplines in the field, it is interesting to measure the diversity of an institution research projects in neuroscience and the level of integration by neuroscientists of theories or methods from different disciplines.
This work is based on various assumptions and technical choices that are questionable.
Based on the Rao-Stirling diversity index, this work proposes a new way to scrutinize interdisciplinary practices. The two components of the overall diversity of sources of the publications of an institution can be interpreted as proxies for the interdisciplinary integrated component of this diversity and its not yet integrated or multidisciplinary component. The decomposition of the overall index as a profile of disciplines meets the demand of users to look inside the overall indicator and to have meaningful and easily interpretable graphical representations. These two decompositions provide institutions with a flexible tool to explore the diversity of disciplines in publication references for each research field. The application of this tool to neuroscience research in French universities shows that it is appropriate to reveal a diversity of practice in terms of interdisciplinarity in a field. These practices could result from an explicit strategy, supported by incentives measures or just emerge in a given institutional setting. In both cases, it is necessary to confirm the interpretation of the quantitative observations provided by these indicators, and further explore the practises to understand which context or policy measures could explain high or low values of the indicators. Applied to all French universities involved in neuroscience research, these indicators make it possible to describe the landscape of interdisciplinary orientation of French neuroscience at the institution level. These indicators could also be used at different levels, as for positioning countries in a global landscape. As other science and technology indicators, these indicators could be successfully used for debating between organisations sharing a common research domain.
Data consist of documents of the four types article, letter, note and review published over the period 2008–2014 in journals classified in the Neurosciences category in the Thomson Reuters database at OST, updated in 2013 with TR updates and with the affiliation validation process carried out by OST with French universities. Documents with less than three references in the WoS are not taken into account.