Why simply summing up any bibliometric indicators does not justify a good composite indicator for individual researcher assessment - A measurement perspective

Compared to single indicators, composite indicators have been little discussed in bibliometrics and scientometrics, and if so, then only very critically, which might be explained by their proximity to the world university rankings which are heavily under attack. A central problem of the use of composite indicators so far is the lack of concepts on how to create composite indicators from individual indicators. Unfortunately, simply summing up any bibliometric indicators does not justify a good composite indicator. A measurement perspective is needed. Using the example of Ioannidis' data, it was possible to show which measurement-theoretical requirements the data must fulfil in order to justify a measurement scale as a composite indicator (one-dimensionality, reliability, invariance, validity, fairness). Although the available data cannot fully meet the quality criteria, in principle the approach of Ioannidis et al. is to be welcomed. The common bibliometric indicators usually cannot even provide information on one of the mentioned quality criteria. Highly reliable and valid composite indicators or scales are required, especially for use in empirical support of individual researcher assessment.