- The paper recalibrates science mapping tools like Pajek and VOSviewer to accommodate the 2011 updates to Web of Science categories and subject areas.
- The updated analysis incorporates 19 factors, improving variance explanation and revealing nuanced disciplinary shifts, particularly in mathematics and computer science.
- This refined methodology enhances the understanding of interdisciplinarity, providing improved tools for research evaluation and policy-making.
Understanding the Evolution of Global Maps of Science with Updated Web-of-Science Categories
The paper "Global Maps of Science based on the new Web-of-Science Categories" by Loet Leydesdorff, Stephen Carley, and Ismael Rafols presents an exploration and recalibration of scientific mapping tools in light of updates made in 2011 to the Web of Science (WoS) Categories. Version 5 of the Science and Social Science Citation Index introduced an adjustment from 222 ISI Subject Categories to 225 WoS Categories, alongside a new set of 151 higher-level "Subject Areas." This restructuring necessitates a re-evaluation of science mapping techniques, especially as they pertain to the validity of global science maps created using these categorizations.
The authors specifically recalibrated their previously developed mapping tools, notably Pajek and VOSviewer, to accommodate the new WoS Categories. Two prominent aspects were addressed: the distinction between WoS Categories and the newly defined Subject Areas, and the reallocation of journals between categories. For instance, journals previously categorized under ISI's "Engineering, Chemical" are now under “Pharmacology & Pharmacy,” illustrating the significant shifts occurring at a granular classification level. The paper underscores that while higher-level aggregate mappings display robustness, discrepancies occur at more detailed levels of aggregation.
A key component of the updated mapping involves the incorporation of 19 factors, as opposed to the previous 18, to explain 54.3% of the variance in the citation matrix. This addition primarily affects mathematics-related disciplines, supporting the argument that mathematics is inherently interdisciplinary, influencing disparate fields like medicine, engineering, and social sciences. The recognition of these connections importantly leads to a split in the "Computer Science" factor into "Computer Science" and "Mathematical Methods," which reflects nuanced shifts in disciplinary acknowledgments.
Moreover, this paper emphasizes the collaborative nature of using tools like Pajek and VOSviewer for visualizing these newly defined clusters. While VOSviewer’s clustering algorithm divides disciplines into four major groups (biomedical, physical, environmental, and social sciences), this organization is simultaneously made available as an optional partition in the Pajek implementation. Despite differences in visualization outcomes between these tools, each adds distinct value in analyzing complex bibliometric data.
The implications of this work are manifold. In practical terms, this updated mapping methodology provides a more accurate lens through which to view interdisciplinarity across scientific domains, enhancing the potential for novel insights in research evaluation and policy-making. Theoretically, these findings urge a reconsideration of traditional disciplinary boundaries, particularly in light of rapidly evolving fields. Speculatively, as AI techniques continue to advance, these refined bibliometric tools could further optimize the structuring of ever-growing scientific repositories, fostering more integrated and accessible networks of knowledge.
In conclusion, the conscientiously redesigned mapping approach introduced in this paper not only accommodates structural changes brought about by WoS updates but also opens new avenues for in-depth analysis of field interrelations. The anticipation is that such refined tools will serve the scientific community better in mapping interdisciplinary research landscapes in the future.