Dimensions was introduced as an alternative bibliometric database to the well-established Web of Science (WoS) and Scopus, however all three databases have fundamental differences in coverage and content, resultant from their owners’ indexation philosophies. In light of these differences, we explore here, using a citation network analysis and assessment of normalized citation impact of overlapping publications indexed in all three databases, whether the three databases offer structurally different perspectives of the bibliometric landscape or if they are essentially homogenous substitutes. Our citation network analysis of core and exclusive 2016–2018 publications revealed a large set of core publications indexed in all three databases that are highly self-referential. In comparison, each database selected a set of exclusive publications that appeared to hold similarly low levels of relevance to the core set and to one another, with slightly more internal communication between exclusive publications in Scopus and Dimensions than WoS. Our comparison of normalized citations for 41,848 publications indexed in all three databases found that German sectors were valuated as more impactful in Scopus and Dimensions compared to WoS, particularly for sectors with an applied research focus. We conclude that the databases do present structurally different perspectives, although Scopus and Dimensions with their additional circle of applied research vary more from the more base research-focused WoS than they do from one another.
Academic data infrastructures facilitate bibliometric research and reporting, as they link bibliometric data with research interests. In Germany, the Kompetenznetzwerk Bibliometrie (KB), a national consortium of 22 institutions, serves this purpose by hosting, probing, and curating large-scale bibliometric raw data from commercial and open providers. Users at partner institutions regularly apply the data infrastructure in workflows. Due to its former focus on commercial data, the KB’s infrastructure has not been detailed transparently to date. However, as an increasing number of bibliometric publications and reports rely upon it, this article provides a comprehensive documentation of the KB’s infrastructure specifically designed for bibliometric applications. It explains the conceptual considerations employed to define the infrastructure’s technical characteristics and procedures. Descriptions are provided regarding (1) the technical infrastructure, (2) the KB’s own database schema consistently applied across several bibliometric databases, (3) processing bibliometric raw data, (4) multi-dimensional quality assurance processes, and (5) additional data curation. By making these procedures transparent and open to scrutiny, we aim to bolster the reliability of our infrastructure and facilitate reproducible research while contributing to the broader discourse on transparency in the field of bibliometrics and providing input to other organisations developing similar infrastructures. Peer Review: https://www.webofscience.com/api/gateway/wos/peer-review/10.1162/qss.a.20
Principal component analysis (PCA) denotes a popular algorithmic technique to dimension reduction and factor extraction. Spatial variants have been proposed to account for the particularities of spatial data, namely spatial heterogeneity and spatial autocorrelation, and we present a novel approach which transfers PCA into the spatio-temporal realm. Our approach, named spatio-temporal principal component analysis (stPCA), allows for dimension reduction in the attribute space while striving to preserve much of the data's variance and maintaining the data's original structure in the spatio-temporal domain. Additionally to spatial autocorrelation stPCA exploits any serial correlation present in the data and consequently takes advantage of all particular features of spatial-temporal data. A simulation study underlines the superior performance of stPCA if compared to the original PCA or its spatial variants and an application on indicators of economic deprivation and urbanism demonstrates its suitability for practical use.
1 December 2013 in Berliner Journal für Soziologie
National bibliometric performance is commonly measured via relative impact indicators which appraise absolute national values through a global environment. Consequently the resulting impact values mirror changes in the national performance as well as in its embedding. In order to assess the importance of the environment in this ratio, we analyse the increase in Chinese publications as an example for a structural change altering the publication universe. Via a counterfactual comparison we observe how Chinese publications differ in their citation-based relevance attribution by benefiting especially Asian countries, while scientifically leading countries in the West gain less than postulated by the former status quo. Due to the substantial quantity of Chinese publications this differentiation in relevance attribution imposes globally noticeable effects on national impact statistics. We argue that such structural changes in the environment have to be taken into account for an unbiased measurement of national bibliometric performance.