Garfield E, Welljams-Dorof A
Institute for Scientific Information, Philadelphia, PA 19104.
Theor Med. 1992 Jun;13(2):117-35. doi: 10.1007/BF02163625.
The purpose of this paper was to determine if quantitative rankings of highly cited research authors confirm Nobel prize awards. Six studies covering different time periods and author sample sizes were reviewed. The number of Nobel laureates at the time each study was published was tabulated, as was the number of high impact authors who later became laureates. The Nobelists and laureates-to-be were also compared with non-Nobelists to see if they differed in terms of impact and productivity. The results indicate that high rankings by citation frequency identify researchers of Nobel class--that is, a small set of authors that includes a high proportion of actual Nobelists and laureates-to-be. Also, the average impact (citations per author) of Nobelists and laureates-to-be is sufficiently high to distinguish them from non-Nobelists in these rankings. In conclusion, a simple, quantitative, and objective algorithm based on citation data can effectively corroborate--and even forecast--a complex, qualitative, and subjective selection process based on human judgement.
本文的目的是确定高被引研究作者的定量排名是否能证实诺贝尔奖的授予情况。回顾了六项涵盖不同时间段和作者样本量的研究。统计了每项研究发表时的诺贝尔奖获得者数量，以及后来成为诺贝尔奖获得者的高影响力作者数量。还将诺贝尔奖获得者和未来的诺贝尔奖获得者与非诺贝尔奖获得者进行了比较，以查看他们在影响力和产出方面是否存在差异。结果表明，按被引频次进行的高排名能够识别出诺贝尔奖级别的研究人员——也就是说，一小部分作者中包括了高比例的实际诺贝尔奖获得者和未来的诺贝尔奖获得者。此外，诺贝尔奖获得者和未来的诺贝尔奖获得者的平均影响力（每位作者的被引次数）足够高，足以在这些排名中将他们与非诺贝尔奖获得者区分开来。总之，基于被引数据的简单、定量且客观的算法能够有效地证实——甚至预测——基于人为判断的复杂、定性且主观的评选过程。
