The comparison of classification-system-based normalization procedures with source normalization alternatives in Waltman and Van Eck (2013)

摘要

Along with other co-authors, Ludo Waltman and Nees Jan van Eck have recently contributed to significantly increasing the fairness of bibliometric research assessments, in particular that of multidisciplinary assessments involving comparisons of citation impact between different fields of science (Van Eck, Waltman, Van Raan, Klautz, & Peul, 2013; Waltman & Van Eck, 2013a; Waltman, Van Eck, Van Leeuwen, & Visser, 2013). This note addresses some pending issues in their latest contribution - Waltman and Van Eck (2013b) for this journal, referred to as WVE hereafter - concerning a systematic large-scale empirical comparison of classification-system-based versus source normalization procedures.1 Among the former, they focus on what we will call classification system WoS, namely, the system where publications are classified into fields based on the journal subject categories in the Web of Science bibliographic database. WVE study the normalization procedure based on this system that uses field mean citations as normalization factors.2 To differentiate this procedure from others of the same type, we denote it by NWoS (rather than NSC as do WVE). On the other hand, according to WVE SNCS exhibits the best performance among the source normalization procedures. Therefore, for our purposes, the issue put forward by WVE is the comparison of NWoS versus SNCS.