The lognormal distribution explains the remarkable pattern documented by characteristic scores and scales in scientometrics
作者：
Highlights：
• CSS produce predictable configurations when applied to particular statistical distributions.
• Typical CSS class configurations can be identified with mathematical tools and statistical analysis of synthetic datasets.
• Application of CSS to lognormal distributions with σ parameter near 1 results in the typical 70–21–6–3% CSS pattern.
• Atypical CSS patterns obtained in previous studies are also explainable as a consequence of lognormal data.
• CSS results are a natural extension of the skewness of science manifest in heavy-tailed distributions.
摘要
•CSS produce predictable configurations when applied to particular statistical distributions.•Typical CSS class configurations can be identified with mathematical tools and statistical analysis of synthetic datasets.•Application of CSS to lognormal distributions with σ parameter near 1 results in the typical 70–21–6–3% CSS pattern.•Atypical CSS patterns obtained in previous studies are also explainable as a consequence of lognormal data.•CSS results are a natural extension of the skewness of science manifest in heavy-tailed distributions.
论文关键词：Citation analysis,Characteristic scores and scales (CSS),Lognormal distribution,Universality claim
论文评审过程：Received 3 November 2017, Revised 7 February 2018, Accepted 7 February 2018, Available online 8 March 2018, Version of Record 8 March 2018.
论文官网地址：https://doi.org/10.1016/j.joi.2018.02.002
