Publica
Hier finden Sie wissenschaftliche Publikationen aus den FraunhoferInstituten. Mean values of skewed distributions in the bibliometric assessment of research units
 Scientometrics 125 (2020), No.2, pp.925935 ISSN: 01389130 ISSN: 15882861 

 English 
 Journal Article 
 Fraunhofer ISI () 
 skewed distribution; mean value; adjusted mean value; Hirschindex; bibliometrics; ranking of research units 
Abstract
Nearly all distributions in bibliometrics are skewed. In particular, the distribution of citations of publications by research units is skewed. In a statistical view, the calculation of mean values can imply misleading or even wrong information. However, in citation analysis, the calculation of mean values of skewed distributions are standard. Therefore, when ranking research units, it is recommended instead to replace the calculation of standard mean values by the calculation of adjusted mean values to exclude outliers with very high citations and those with very few or no citations as well. Such an adjusted mean value is oriented on the standard activity of a research unit and results in a more adequate assessment. This approach is based on the Hirschindex concept. The calculation results in a different ranking of research units, which may be important in cases where the distribution of finances depends on bibliometric rankings. In addition, a differentiation between standard activities and excellent results is possible, thus opening two dimensions of the assessment of research units.