Insights into the accuracy of the World’s Top 2% of Scientists list by Stanford University

The tools and methods used to measure the impact of an individual’s research work are H-index, citations, Altmetrics, and impact factor of journals where the work is published. The databases widely used for the H-index and citations are Web of Science, Scopus, and Google Scholar. Web of Science indexes over 10000 journals while Scopus indexed over 15,000 journals. Google Scholar has more journals indexed and more publication types than the other databases but it is not comprehensive and accurate, because individual researchers determine the results, unlike the other two. The h-index is considered a measure of both the scientific productivity and scientific impact of a scientist. In citation analysis, the number of times an article is cited by other works is used to measure the impact of a publication/author. Altmetrics is based on the attention of a published work through social media, citations, and article downloads. Impact factor (IF) is a measure of the importance or rank of a journal, it is based on the citations published papers the previous two years.
A few years ago John Ioannidis and co-authors of Stanford University created a unique publicly available database of top-cited scientists in the world using Scopus data from Elsevier, which has enthralled the scientific community, institutions, and media[1-3]. This database, touted as a new approach to eliminate the misuse of citation metrics, has enthralled the scientific community, institutions, and media. Many institutions used this as a yardstick to assess the quality of researchers, some institutions widely publicized the list and gave rewards to researchers who were listed. At the same time, some researchers look at this list with skepticism by citing problems with the methodology used in the c-score-based ranking. To evaluate the accuracy of c-score-based ranking, being my name figured in both the career and individual year lists, I have done a detailed analysis of the matrix parameters. 2578 and 4635 Indian authors appear in the career and single-year categories, respectively. For comparison of the ranking, the last 25 years (1998-2022) of Nobel laureates of Physics, chemistry, and medicine, and the top 100 rank holders in the list are chosen. The latest career-long, and single-year-based databases (2022) were used for this analysis. The details of the analysis are presented here. 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 [1]. The single-year database contains 210199 names. So, the list contains names of the top 4 %, if we assume 100000 scientists on the list as 2%! The entry of many authors having single digit H index and a very meager total number of citations indicates serious shortcomings of the c-score-based ranking methodology.

衡量个人研究成果影响之工具与方法，包括H指数、引用次数、Altmetrics以及发表作品之期刊影响因子。广泛使用的H指数与引用次数数据库包括Web of Science、Scopus与Google Scholar。Web of Science收录超过10,000种期刊，而Scopus则收录超过15,000种。相较于其他数据库，Google Scholar收录的期刊与出版类型更为丰富，但其全面性与准确性不足，因其结果由个别研究者决定，与前者不同。h指数被视为衡量科学家科学生产力和科学影响力的指标。在引用分析中，文章被其他作品引用的次数被用来衡量出版物/作者的影响力。Altmetrics基于发表作品在社会媒体、引用次数与文章下载量上的关注度。影响因子（IF）是衡量期刊重要性或排名的指标，其基于前两年发表论文的引用次数。数年前，斯坦福大学的John Ioannidis及其合作者利用Elsevier的Scopus数据，创建了一个独特的、公开可用的全球顶级引用科学家数据库，该数据库引起了科学界、机构与媒体的极大关注[1-3]。该数据库被誉为消除引用计量指标误用的新方法，引起了科学界、机构与媒体的极大关注。许多机构以此作为评估研究者质量的标尺，一些机构广泛宣传该名单，并对列入名单的研究者给予奖励。与此同时，一些研究者对此名单持怀疑态度，指出基于c-score排名的方法存在方法论问题。为了评估基于c-score排名的准确性，鉴于本人姓名同时出现在职业与个人年度名单中，我对矩阵参数进行了详细分析。在职业与单一年度类别中，分别有2578名和4635名印度作者出现。为了比较排名，选取了1998年至2022年（最后25年）的诺贝尔物理学、化学与医学奖得主以及名单上排名前100位的人选。本次分析使用了最新的职业长期与单一年度数据库（2022年）。分析细节如下。尽管文章称选择基于c-score（含自我引用）前10万位科学家或子领域中2%或以上的百分位排名，但实际基于职业的排名名单包含204,644个名字[1]。单一年度数据库包含210,199个名字。因此，名单包含了排名前4%的科学家，如果我们假设名单上有100,000位科学家，则2%即为4%。许多作者H指数个位数且总引用次数极为有限的情况，表明基于c-score排名的方法存在严重缺陷。