**2.3. Use of altmetrics in science evaluations**

With regard to the practical application of altmetrics in research policy, science evaluations, and management, the scientific community is mostly skeptical. Bornmann and Haunschild [13] stress the problematic nature of the matter, namely that altmetrics should first confirm with the Leiden Manifesto for research metrics [14] before being applied on a greater scale. The central difficulties associated with altmetrics are presented, namely that there are currently no standardized indicators, that altmetric data are for the most part not accessible in a transparent and open manner, and that numbers can be manipulated through "gaming." Gaming is a term for the targeted manipulation of data for the purposes of achieving better altmetric values. Such gaming activities are negative side effects of an orientation along user statistics in evaluation practice [9]. However, in spite of the difficulty in consistently unambiguously distinguishing gaming from marketing, altmetrics service providers are trying to minimize such effects. For example, altmetric.com manually removes obvious manipulations of altmetric scores or limits them by means of spammer lists [15].

Gaming is also a problem beyond the sources assessed by altmetric service providers. In a study by Meier and Tunger, it became apparent that it is possible to considerably influence the metrics specially developed by the ResearchGate platform, the RG score [16]. The RG score is intended to measure the "scientific reputation" of ResearchGate users. It is influenced by the impact of a user's own scientific publications but also by their social activities on the platform (see https://www.researchgate.net/RGScore/FAQ). Meier and Tunger found that it is possible within a relatively short time to achieve an RG score that is higher that the RG scores of half of all RG users solely by gaming without any scientific publications.

In another study for the European commission, Kim Holmberg found that altmetrics are not yet practically applied in the EU for the purposes of scientific evaluation. In his view, such practice on a wide scale would be premature as long as what altmetrics actually measure remains unclear [17].
