Integrating multiple sources of information in literature-based maps of science

Abstract

The last decades have seen an increasing utilization of graphical representations of aspects of science. These "maps" are mainly used to depict the underlying relational structures of publications, or publishing entities, within the science and technology system. To this end, the maps generally draw on only one source of bibliometric (i.e., literature-based quantita tive) data—mostly citations or keywords. These information items will necessarily descnbe only one facet of the (intellect ual and/or social) structure of science. We argue that in order to obtain a more complete description of the common underly ing structure one requires the incorporation of more sources of (bibliometric) data. We discuss the statistical model INDSCAL that combines multiple sources (e.g., citations, keywords, sub ject classificanon codes). This model yields an integrated multi-dimensional representation of the relational structure of a set of scientific entities common to all sources. The utility of INDSCAL is illustrated in an application to core journals in the field of astronomy and astrophysics. Some marked dif ferences occur in the journal structure when based on more data sources than citations alone.