Scientometric and network analyses
Authorship and citation data were analyzed using descriptive statistics,
scientometric indices, and network analysis methods \citep{Nita_2019a,Barab_si_2002,Aria_2020}. For scientometric
data, we considered the following metrics: the number of articles published by an author, the number
of citations received by hierarchical modeling papers, the local h-index of
an author (indicating the number of publications for which an author has been cited by
articles in our database at least that same number of times), and the number
of citations from our database received by the papers referenced in the
Literature Cited section (i.e., papers included in the reference
sections of analyzed hierarchical modeling papers)\citep{Aria_2017}.
To understand co-authorship and co-citation patterns, we used network
analysis \citep{borgatti2018analyzing}. We generated two undirected, unweighted networks: (1) a
co-authorship network, where authors (nodes) are linked to other
authors (edges) if they share at least one coauthor (including
themselves), and (2) a co-citation (co-references) network, where paper references
(nodes) relate to other paper references (edges) if they share
at least one reference (including the references that serve as nodes).
The co-authorship network was used to identify network leaders, while
the co-citation network highlighted the most important papers referenced
in the field \citep{RN770}. For each of the two networks, we also calculated two
node-level centrality metrics: degree and normalized
betweenness \citep{borgatti2018analyzing}. Degree centrality of an author represents the number of
direct connections that the author has with other authors in the
network and helps to identify the most collaborative authors (i.e., the
authors with the highest number of connections). Betweenness
centrality measures the extent to which an author lies on paths between
other authors in the network \citep{Nita_2019a}. Such authors may be considered ”bridge”
authors (i.e., they have the potential to influence the research topics in a co-authorship network) \citep{borgatti2018analyzing,Nita_2019a}. Co-authorship and co-citation networks were calculated
using VosViewer \citep{RN770} and graphically represented
using the NodeXL app \citep{RN768}. Node-level metrics were calculated
using the R igraph package \citep{RN769}. Scientometric indices were extracted via bibliometrix R package \citep{Aria_2017}.