摘要：A critical discussion is presented for the Author Metrics Database (AMD) created by Ioannides et al. (2016, 2020) containing citation-based indicators for 165,000 authors publishing in journals indexed in Scopus. It is concluded that the AMD is a rich intermediary dataset open for further analysis to all interested users. However, its indicators suggest a false precision and lack transparency. The theoretical and statistical basis of the database’s key composite impact indicator is weak, and information on whether or not underlying author publication lists were validated is lacking. The paper aims to broaden the perspective on the further development of an AMD, highlighting its bottom-up, interactive use, aptness for self-assessment and educational function for a wide user community. POLICY HIGHLIGHTS Scopus diverges from Eugene Garfield’s original concept of the Science Citation Index, as citation impact plays a weaker role as journal selection criterion. The transparency of the Article Metrics Database (AMD) is seriously hampered by the lack of information on whether the data were verified by scientists themselves. A complex composite indicator in the AMD decides whether or not a particular author is included. Its components are strongly statistically dependent and are largely based on the position an author has in a paper’s author sequence but lack a sound theoretical foundation. An assessment of an individual researcher cannot be merely based on whether or not he or she is included in the AMD. The issue as to how to deal with multi-authored papers in research assessment of individuals can to some extent be enlightened by bibliometric indicators but cannot be solved bibliometrically. This is why the Composite Indicator suggests a false precision. The AMD focuses almost exclusively on senior scientists. Early career scientists and emerging research groups who will shape science and scholarship in the near future hardly appear in the AMD. Desktop bibliometrics using the AMD as a sole source of information must be rejected. Using the AMD as a starting point in a more extensive bibliometric data collection makes it de facto a promotion tool for other Elsevier products. An alternative approach is an interactive, bottom-up bibliometric tool designed for self-assessment and educational purposes, showing how bibliometric indicators depend upon the way in which initial publication lists, author benchmark sets, subject delimitations, thresholds and evaluative assumptions are chosen. Research assessment is much more than just bibliometrics. It requires an overarching evaluative framework based on normative views on what constitutes research performance and which policy objectives should be achieved.
