Abstract:General interpretationInterpretation in the context of interregional knowledge networks
Examples of applications to interregional knowledge networks
Degree centralityLocal measure of node centrality and a node's embeddedness in a network
Note:The number of link changes counts the number of ceased and newly created linkages. For a graphical overview see also Figure A2.
“…Salman and Saives (2005) indicate that centrality variables tend to be correlated to one another, which could result in a multicollinearity of variables. Therefore, we choose degree centrality, which is widely discussed (Coffano, Foray, & Pezzoni, 2017;Mitze & Strotebeck, 2018;Tsai, 2001;Wang, Rodan, Fruin, & Xu, 2014). Degree centrality is a classic measure of network position, which refers to the number of other nodes directly connected to one node (Freeman, 1979).…”
Section: Degree Centrality and Knowledge Outputmentioning
There is consensus among scholars that social networks are important mechanisms of knowledge spillovers. By occupying a central and advantageous network position in the collaboration network, it is easy for actors to access intangible external resources. However, empirical studies of the impact of network structure on knowledge production remain scarce. Based on copublication data from the Web of Science database (WoS) from 2000 to 2015, we construct eleven international scientific collaboration networks and empirically estimate the impacts of multiple network properties, comprehensively measured by degree centrality, structure holes, and small‐world quotient, on national knowledge output. Empirical results based on fixed effect negative binomial models suggest positive effects of the three facets of network properties, that is, higher degree centrality, structural holes, and small‐world quotient are beneficial for facilitating and improving national knowledge production, which in turn encourages international academic collaboration.
“…Salman and Saives (2005) indicate that centrality variables tend to be correlated to one another, which could result in a multicollinearity of variables. Therefore, we choose degree centrality, which is widely discussed (Coffano, Foray, & Pezzoni, 2017;Mitze & Strotebeck, 2018;Tsai, 2001;Wang, Rodan, Fruin, & Xu, 2014). Degree centrality is a classic measure of network position, which refers to the number of other nodes directly connected to one node (Freeman, 1979).…”
Section: Degree Centrality and Knowledge Outputmentioning
There is consensus among scholars that social networks are important mechanisms of knowledge spillovers. By occupying a central and advantageous network position in the collaboration network, it is easy for actors to access intangible external resources. However, empirical studies of the impact of network structure on knowledge production remain scarce. Based on copublication data from the Web of Science database (WoS) from 2000 to 2015, we construct eleven international scientific collaboration networks and empirically estimate the impacts of multiple network properties, comprehensively measured by degree centrality, structure holes, and small‐world quotient, on national knowledge output. Empirical results based on fixed effect negative binomial models suggest positive effects of the three facets of network properties, that is, higher degree centrality, structural holes, and small‐world quotient are beneficial for facilitating and improving national knowledge production, which in turn encourages international academic collaboration.
“…In a final step, further regional covariates are added to the regionally aggregated data set on research collaborations. This extension enables applied researchers to conduct multivariate analyses, for instance, in order to investigate the regional determinants of network formation within the boundaries of a closed national research network (see, for instance, [1] for the estimation of a gravity-type model of network formation in knowledge networks as well as [2] for a related study analyzing the determinants of a region׳s centrality position in research collaboration networks and its variation over time). Regional covariates have thereby been selected as to represent node properties and the relationship between nodes [9] .…”
Section: Experimental Design Materials and Methodsmentioning
confidence: 99%
“…Details on the evolution of the German biotech industry, the formation of BioRegions and the role of public funding therein are provided in the supplementary materials of this article. Further information can also be found in [1] , [2] and [13] . Fig.…”
Section: Experimental Design Materials and Methodsmentioning
confidence: 99%
“…2 suggests that interregional research collaborations are stable over time, an early analysis [2] has shown that the number and direction of research collaborations (both at the individual and regional levels) have changed significantly between 2005 and 2009. This implies that the data may also be used to analyse the short-term evolution of knowledge networks as done in [2] .…”
Section: Experimental Design Materials and Methodsmentioning
confidence: 99%
“…Although the data are currently organized in a cross-sectional manner, the data set can be easily extended to a panel structure as done in [2] for a subset of variables presented here. Moreover, the BIOCOM industry directory is published on an annual basis, which allows applied researchers to further extend the sample range beyond the two yearly snapshots presented here.…”
Section: Experimental Design Materials and Methodsmentioning
This article describes a data set to map and model research collaborations in German biotechnology. Underlying micro-data for firms and institutions in the biotech sector together with information on their research collaboration partners have been extracted from a commercial industry directory, the BIOCOM Year and Address book, for 2005 and 2009. The data have been processed and aggregated to the level of NUTS3 regions. This core data set has been linked to regional covariates which measure the regional endowment with biotech-related research capacities, sector-specific S&T policy support and the strength of a region׳s overall local innovation system. The full data set, which is attached to this article, offers applied researchers an alternative source of information for empirical analyses of knowledge flows in research networks and for studying their determinants. Potential fields of application include social network and regression analysis. First empirical results are reported in “Determining factors of interregional research collaboration in Germany׳s biotech network: Capacity, proximity, policy?” (Mitze and Strotebeck, 2018) and “Centrality and get-richer mechanisms in interregional knowledge networks” (Mitze and Strotebeck, 2018).
This paper investigates the role played by the position of European regions in research networks on their rate of innovation and economic growth. The analysis is based on a panel of EU-28 NUTS2 regions participating in EU Framework Programmes observed over the 2004-2014 period. We find that regions that are more central in the network (higher strength centrality) and those that are surrounded by highly inter-connected regions (higher clustering index) show higher rates of innovation and higher economic growth. We also find heterogeneous effects of centrality and clustering for peripheral and central regions. We conclude that a more interconnected network (an increase in centrality for peripheral regions and of clustering for urban areas) would create benefits both at the periphery and at the core of Europe.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
