A Study on the Applicability of Author Identification Numbers in Scientific and Technical Paper Databases

Subjects: Management Science >> Science ology and Management Subjects: Library Science，Information Science >> Information Processing submitted time 2024-06-06

Dongbo Shi Hui Deng Zhijian Yang Ningjie Yang Yuxiu Liu Yufei Mao

Abstract： Purpose To evaluate the coverage and accuracy of author identification number (author ID) of the major bibliographic databases and to assess whether they could be directly used in empirical research.
Methods The ground truth data set consists of articles from 825 Chinese scientists. The coverage, accuracy, and robustness of each author ID are calculated by retrieving and collecting the IDs of scientists and their respective publication information in the bibliographic databases. The validity of the author IDs for empirical research is assessed by replicating a top journal empirical article using the data collected through author IDs.
Results First, WOS, Scopus, AMiner, and OpenAlex can retrieve more than 90% of Chinese scientists’ identifiers, while ORCID’s coverage is less than 50%. Second, the accuracy of Scopus is the highest at 85.2%, and the accuracy of OpenAlex is the lowest at only 51.2%. Third, directly using the publication data collected through author IDs for empirical research will introduce non-negligible bias.
Limitations The ground truth data set is limited, because it is mainly composed of young scientists, and lack scientists from social sciences and humanities.
Conclusion At present, the author identification number of the major databases cannot be directly applied to the empirical research of large-scale data. A standardized information platform for scientists’ publications is needed to overcome the author-name disambiguation problem.

Peer Review Status:Awaiting Review

 Hits 599  Downloads 124  Comment 0