AIM: A plethora of journal indicators exists, and are used e.g. in national assessment exercises as part of the basis for university funding. Recent studies report as many as 39 indicators of which the major part are citation based. While studies have shown these citation-based journal indicators to correlate, it is assumed that some indicators may be better at discriminating “good” from “bad” journals, and it is consequently assumed that some indicators are more robust than other. This study aims at describing these attributes for a selection of the most promising and/or most used citation-based journal indicators. METHODS: We use Bayes' theorem to illustrate the connection between journal indicators and the underlying citations. This method is closely related to that of Lehmann, Jackson & Lautrup (2006). This is a work in progress. MATERIALS: Complete citation data from Web of Science for select subject categories. As this is a work in progress, subject and indicator selection is not yet complete, nor is data acquisition. RESULTS: Examplar data may be seen in figure 1, displaying the robustness of the journal impact factor for describing citations. REFERENCES: Lehmann, S., Jackson, A.D., Lautrup, B. (2006). Measures for Measures. Nature, 444:1003-1006Figure: Robustness of the journal impact factor.
Main Research Area:
16th Nordic Workshop on Bibliometrics and Research Policy, 2011