Author name disambiguation: What difference does it make in author-based citation analysis?

In this paper, we explore how strongly author name disambiguation (AND) affects the results of an author-based citation analysis study, and identify conditions under which the commonly used simplified approach of using surnames and first initials may suffice in practice. We compare author citation ranking and co-citation mapping results in the stem cell research field 2004-2009 between two AND approaches: the traditional simplified approach of using author surnames and first initials, and a sophisticated algorithmic approach. We find that the traditional approach leads to extremely distorted rankings and substantially distorted mappings of authors in this field when based on firstor all-author citation counting, whereas last-author based citation ranking and co-citation mapping both appear relatively immune to the author name ambiguity problem. This is largely because romanized names of Chinese and Korean authors, who are very active in this field, are extremely ambiguous, but few of these researchers consistently publish as last authors in by-lines. We conclude that more earnest effort is required to deal with the author name ambiguity problem in both citation analysis and information retrieval, especially given the current trend towards globalization. In the stem cell field, where lab heads are traditionally listed as last authors in by-lines, last-author based citation ranking and co-citation mapping using the traditional simple approach to author name disambiguation may serve as a simple workaround, but likely at the price of largely filtering out Chinese and Korean contributions to the field as well as important contributions by young researchers.