Subject-access metadata on ETD supplied by authors: A case study about keywords, titles and abstracts in a Brazilian academic repository

Abstract

The keyword lists are rich in terminology and as such, they are also characterized by great semantic ambiguity, which presents the problems of synonymy and polysemy typical of uncontrolled language, which is both an advantage and a drawback. However, assigning keywords has gained particular importance in the current open scientific communication ecosystem in digital environments, mainly in the academic dissertation and repositories of theses. This article reports a study on the organization of knowledge, using subject-access metadata of master's theses from the digital repository of the University of São Paulo (USP), Brazil, applying analytical techniques to a big amount of data. The objective was to analyze the number of keywords in each record and keywords repeated in the title and abstracts of the dissertations in Portuguese and English. The analysis of the 48,501 metadata records of master's theses submitted to the repository, between 2001 and 2019, presents a total of 223,867 keywords in Portuguese, with an average of 4.62 keywords per record, and a total of 216,521 keywords in English, equivalent to an average of 4.59 per record. Although, the attribution of keywords in Portuguese and English by the author is an economic way to expand access to the content of theses and dissertations in institutional repositories, it is necessary to define rules for authors about the choice of keywords and the preparation of the abstract, as well as its translation into English. The task of controlling keywords requires a partnership between authors and librarians who can enrich the quality of indexing languages.