Figure 1.
Example of bibliographic coupling network and co-authorship network stemming from a set of three articles.
Table 1.
Keywords.
Figure 2.
Mapping the process of data analysis.
Figure 3.
Conceptualization of the research questions.
Starting from the article i written by α,β and γ and published in year t, two networks are derived. On the left side, the co-authorship network shows that α is already part of a group of other authors and had joint works with two of them (represented by the two edges inside the author set). β and γ are connected to the set of authors through α by means of their new joint-work i. We then extract three measures, degree, closeness and betweenness for each co-author of i, and map only the maximum value into i. On the right side, the article i have common references with articles of the sets A and B. We measure the degree, betweenness centralities and clustering coefficient of the article i. We construct the dependent variable cumulating the citations received by article i in the two years following the year of its publication. Last, we verify the effects the co-authorship and bibliographic coupling networks measures on the citation count of article i.
Table 2.
Spearman correlation coefficients with significance level.
Figure 4.
The expansion of the Vulnerability literature: number of authors (left) and publications (right) per year.
On the left part of the figure, the dark line represents the cumulative number of authors who at least have one publication in the dataset, while the light line represents the number of new authors.
Table 3.
Results of regression analysis.
Table 4.
Number of authors in the Giant Component (GC) with respect to the total authors in the network.
Figure 5.
The evolution of the Vulnerability literature with the Bibliographic Coupling network.
Nodes are represented by articles in the dataset and the edges link articles which share one or more references. In this network, we show only articles with at least 10 citations and that are connected to the giant component (1164 nodes in 2010 that are 72,48% of amount of articles with more than 10 citations).