Abstract

Journals are ranked in the subject categories of the Journal Citation Reports (JCR) according to their impact factors (JIF). The calculation of this indicator includes both citations received from other journals and self-citations from the journal itself. In this paper the current model of the JCR is compared with an alternate model using the JIF without self-citations as a variable. In order to do this, simulations were carried out using JIF without self-citations as an alternative variable. We studied the relationship between the two variables (JIF with self-citations and JIF without self-citations), the gain obtained by the journals with the usual JIF with self-citations, the changes in the ranking in the whole JCR, and changes in the internal ranking in the JCR’s subject categories. A small negative correlation was found between the gain that the journals obtained and their JIF. The results obtained suggest that if JIF without self-citations were used as the variable, the ranking of journals in the JCR and within the subject categories would be, in general, very different from the current one. The quantification of these differences is one of the more interesting results of this work.

Abstract

Citation analysis can provide us with models of the evolutionary dynamics in scholarly and scientific communication. We propose to distinguish between institutional research evaluation (usually, ex post) and knowledge evaluation ex ante, in relation to directionality in citation analysis. We discuss the theoretical literature on communication systems which distinguishes between information and meaning, in which the concept of redundancy plays an important role as measure of the potential of a communication system. This is the basis for a model of knowledge dynamics which differentiates between observable variation and latent selection environments. We use indicators at the journal level and analyze the citation environments of journals in both the cited and citing directions. Among journals, the citing direction can be analyzed by co-citation and indicates the integration of knowledge from different fields. The cited direction can analogously be analyzed by bibliographic coupling and represents the extent to which the cited journal has become relevant for different disciplines, hence indicates knowledge diffusion. We apply this analysis on three different case studies of journal-journal relations: a small scale study of the journal Public Understanding of Science, a random sample of 100 journals, and a large-scale analysis of the set of JCR 2016 journals. Combined, the results seem to confirm the hypothesis that interdisciplinarity cannot be captured by one-dimensional citation analysis. Both citing and cited directions are relevant for knowledge and research evaluations, respectively. We raise the question whether indicators of interdisciplinarity can be developed by combining both directions in citation analysis, indicate further research, and discuss the normative implications of our preliminary results.