The present paper deals with network analysis of collaboration patterns in Statistics focusing on co-authorship relations. Attention to this discipline derives from several motivations. Unlike other fields, co-authorship behavior in Statistics has not yet been investigated. This discipline presents some characteristics common to natural sciences as well as social sciences, and it plays a central role in addressing problems in everyday applications. Therefore, it is of interest to examine what are the network properties and the emerging patterns characterizing this discipline as well as the effects collaboration has on individual performance.
To this purpose, we collected co-authorship data on the 792 academic statisticians in Italy as recorded in the Italian Ministry of University and Research database in March 2010. We used three bibliographic archives including both top-international as well as nationally oriented publications: ISI-WoS, Current Index to Statistics, and bibliographic information related to nationally funded research projects.
Given that each data source showed peculiar characteristics affecting network results (as reported in De Stefano et al. 2013), in this paper we aim at merging the three databases to obtain an unified archive and to use it as a new basis for network analysis. Specifically, two main challenges are managed to obtain the unified co-authorship network: how to combine information from heterogeneous sources by identifying duplicate records, and how to deal with issues related to authors synonyms and homonymies (i.e. name disambiguation) to guarantee the data quality. The merged co-authorship network will be used to describe collaborative behaviour among Italian statisticians in a comparative way with respect to the previous findings based on network analysis of each data sources.
