Disaggregated research evaluation through median-based characteristic scores and scales: a comparison with the mean-based approach
作者：
Highlights：
• Median-based CSS are discussed as an alternative to mean-based CSS.
• Under median-based CSS the class of poorly cited papers decreases; the others are enlarged.
• Moving from mean to median-based CSS leads to a “Matthew effect” at the level of journals.
• Disaggregated research evaluation is favored against the use of unique aggregated indicators.
摘要
•Median-based CSS are discussed as an alternative to mean-based CSS.•Under median-based CSS the class of poorly cited papers decreases; the others are enlarged.•Moving from mean to median-based CSS leads to a “Matthew effect” at the level of journals.•Disaggregated research evaluation is favored against the use of unique aggregated indicators.
论文关键词：Characteristic scores and scales (CSS),Scientometrics,Research evaluation,Disaggregation
论文评审过程：Received 11 March 2017, Revised 17 April 2017, Accepted 17 April 2017, Available online 29 June 2017, Version of Record 29 June 2017.
论文官网地址：https://doi.org/10.1016/j.joi.2017.04.003
