The proposal of using scaling for calculating field-normalized citation scores

Since the end of the 1980s, citation impact values – especially for evaluative purposes – are increasingly presented as mean normalized citation scores (MNCS) than as bare citation counts or citation rates. In rather popular variants of the MNCS the average MNCS over a publication year is not exactly one due to multiple Web of Science subject categories per paper. We propose a scaling method which introduces slight changes in the MNCS values of each paper but ensures that the average value of all MNCS values equals one.