Does C-score-based world ranking give an accurate account of the impact of research papers?

John Ioannidis et. al [1] created a publicly available database of top-cited scientists in the world (popularly known as the world's top 2% Scientist list of Stanford University). This database has generated a lot of interest among the scientific community, institutions, and media attention. Many institutions gave wide publicity to this ranking and prominently displayed the names of their colleagues in various media. Some institutions even rewarded those who made the list in different ways. At the same time, many people look at this list with skepticism, citing the approach and shortfalls in the formula used in the evaluation of the c-score. This database is based on a unique composite indicator (c-score), where standardized details such as citations, h-index, co-authorship adjusted hm-index, and citations to papers in different authorship positions are used[1]. The original database has been updated yearly since its creation in 2019[2-3]. Two separate databases are created based on career-long and, single recent year impact. Career-long data is based on the citations from 1996 to 2022, but the single-year database is based on citations received during the calendar year 2022. This database is created using Scopus data from Elsevier. The latest database is performed using all Scopus author profiles as of October 1, 2023[3]. The Scientists included in this database are classified into 22 scientific fields and 174 sub-fields. To address the concerns originating from the research community on the authenticity of this database, and to assess the strengths and weaknesses of the c-score-based evaluation, I have done a detailed analysis of the matrix parameters of the last 10-year Nobel laureates of Physics, chemistry, and medicine. In this analysis, the latest career-long database is used. In the case of Physics Nobel laureates, analysis is done with the details of Nobel laureates of the last 25 years. The details of the analysis are presented in this article. Though the article says the selection is based on the top 100,000 scientists by c-score (with and without self-citations) or a percentile rank of 2% or above in the sub-field, the actual career-based ranking list has 204644 names. In the career-based rank list, the person with the lowest h index in the list has an h22 of 2 (with an np6022 of 11 and total citations nc9622 of 835). There are 1155 authors with single-digit h22 index too. In the same list, the person with the lowest number of papers has an nc6022 of 2 (with nc9622 of 12128 and first and last publications in 1929 and 1966, respectively). The author with a minimum number of citations in the list has an nc9622 of 41 (h22 of 3). The c-score-based lowest rank holder in the list has a rank of 4809825 (nc9622 of 41, and h22 of 3). These results indicate aberrations in the ranking.

John Ioannidis 等人[1]构建了一个公开可用的全球高被引科学家数据库，即广为人知的斯坦福大学全球前2%科学家榜单。该数据库在科学界、科研机构中引发广泛关注，同时获得了诸多媒体报道。诸多机构对该榜单进行了大力宣传，在各类媒体上醒目展示其上榜同事的姓名，部分机构还以多种形式对上榜者予以奖励。与此同时，不少人对该榜单持怀疑态度，质疑其用于评估c-score的方法与公式存在的缺陷。
该数据库基于独特的复合指标（c-score），综合了被引频次、h指数（h-index）、合作者调整后的hm指数（co-authorship adjusted hm-index），以及不同作者署名位置论文的被引情况等标准化细节[1]。
该原始数据库自2019年创建以来每年更新[2-3]。其分为两个独立数据集：其一基于科研生涯长期学术影响力，其二基于单年度学术影响力。长期影响力数据集以1996年至2022年的被引数据为基础，而单年度数据集则以2022日历年期间获得的被引数据为依据。
本数据库依托爱思唯尔（Elsevier）旗下的Scopus数据库构建。最新版数据库整合了截至2023年10月1日的全部Scopus作者档案[3]。
纳入该数据库的科学家被划分为22个科学领域及174个细分领域。
为回应科研界对该数据库真实性的质疑，并评估基于c-score的评估方法的优劣，笔者针对近10年物理学、化学及医学领域的诺贝尔奖得主的矩阵参数展开了详细分析。本次分析采用了最新版的长期影响力数据集。针对物理学诺贝尔奖得主，分析涵盖了近25年的诺奖得主相关数据。本文将呈现该分析的详细结果。
尽管文中提及遴选标准为按c-score（包含自引（self-citations）与排除自引两种情况）排名前100000位的科学家，或在细分领域中百分位排名（percentile rank）处于前2%及以上的学者，但实际的生涯排名榜单共包含204644个姓名。
在该生涯排名榜单中，上榜者的最低h指数为h22=2（对应np6022=11，总被引频次nc9622=835），另有1155位作者的h22指数为个位数。
同一份榜单中，论文数量最少的作者的nc6022=2（其nc9622=12128，首次与最后发表论文分别为1929年和1966年）。
榜单中总被引频次最低的作者的nc9622=41（h22=3）。
该榜单中基于c-score的最低排名者为4809825位（nc9622=41，h22=3）。
上述结果表明该排名存在异常之处。