The Functional Requirements for Bibliographic Records (FRBR) Framework for Indexing Monographs: Implications for the Book Citation Index™ and Metric Evaluations

In the past, bibliographic data and citation data pertaining to books were difficult to retrieve. Now, as digital resources have improved, so has the priority to advance book-related metrics. This is partly due to the introduction of Thomson Reuter’s Book Citation Index (BKCI) (Adams & Testa, 2011) and the addition of books to Elsevier’s Scopus. These commercial databases; however, are not the ‘be-all and end-all’ for the discerning bibliometrician. Recent assessments of the BKCI (in particular) point to numerous indexing problems, which can lead to flawed evaluations (Gorraiz et al., 2013; Leydesdorff & Felt, 2013; Torres-Salinas et al., 2014). Still, researchers continue to use the BKCI or Scopus, and work mainly with book citations from journal articles (Hammarfelt, 2011; Zuccala et al., 2014), or choose alternative resources, like Google Books (Kousha & Thelwall, 2009), Google Scholar (Kousha & Thelwall, 2011) and OCLC WorldCat (Torres-Salinas & Moed, 2009; White et al., 2009). Concerted efforts are even being made to assess data that has been retrieved from multiple resources (e.g., Zuccala & Cornacchia, 2016). The bibliometrics community is making rapid progress, but there are still several issues that need to be addressed. One in particular is central to all previous studies combined: regardless of where and how bibliographic and citation data are collected, it is essential to recognize that books often belong to bibliographic ‘families’. Since ‘bibliographic families’ can be examined both theoretically and empirically, the aim of our study is to examine and explain several interrelated concepts linked to a family-oriented entity-relationship model, known as the Functional Requirements for Bibliographic Records (FRBR). Here, we have chosen to use this model to illustrate the extent to which books, as complex entities, are not always indexed accurately with appropriate metadata. In the second part of our study, we will present some data collected and assessed from the BKCI, OCLC-WorldCat, and Goodreads, and use this data to demonstrate why a robust model is necessary, first for the practice of indexing books, and more critically for the practice of book-oriented metrics. The empirical aspect of our research is based on the following question: Do books currently indexed in the Book Citation Index have adequate metadata and/or data designed to reflect inherent familial components and relationships?