Aplicação de ciência de dados na análise do posicionamento autoral e contribuições científicas em artigos

• Autor(a) principal: Souza, Edson Melo de
• Data de Publicação: 2021
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações da Uninove
• Texto Completo: http://bibliotecatede.uninove.br/handle/tede/2788
• topic: ciência de dados; posição autoral; lista de contribuições; byline; data science; author position; contribution list; byline; CIENCIA DA COMPUTACAO::SISTEMAS DE COMPUTACAO

In recent years, there has been an increase in the number of authors listed in articles and the issue of authorship and scientific contributions is still open in the literature since there is no formal definition or recommendation for the authors’ position. Objective: In this dissertation, I’ve scrutinized the relationship between scientific contribution and authorship ethical issues. My main objective was to study contributions types and its impact in authorship position. Methodology: I’ve collected 2, 024 articles containing 17, 385 authors from the following data sources: SCImago, Scopus, and the following jour- nals: Annals of Internal Medicine (Ann. Intern. Med.), Journal of the American Medical Association (JAMA), and PLoS Medicine (PLoS Med.). All of those being journals from the biological sciences and medicine areas. To analyze the data I’ve employed statisti- cal techniques, specifically Factor Analysis; and also data science techniques, specifically Linear Regression. Furthermore, for data collection and cleaning automation, I’ve deve- loped robots with data scraping techniques. Results: My study revealed the existence of mainly two scientific contribution categories (“Theory” and “Methodology/Logistics” ). I’ve also proposed an universal categorical model of scientific contribution based on th- ree main categories: theory, methodology, and logistic. Furthermore, I’ve also developed algorithms to automate data collection by robots/crawler. Conclusion: Data science te- chniques have allowed the construction of robots to automate data collection in scientific databases and publications together with developed algorithms, substantially reducing the collection time and improving the accuracy of the processes. The grouping of sci- entific contributions showed that the largest contributions are found in the theoretical group, indicating that the academic experience of the authors is a major factor, while the methodological and logistical contributions represent substantial contributions. The categorical model proposed for the study of the relationship between authorial positioning and scientific contributions shows that authors who contribute theoretically tend to be the first authors. Logistic contributions, on the other hand, tend to place an author last. Methodological contributions do not have an impact on author position. Finally, the small effect registered in the correlation between the contribution variables evidences that the author’s position is not defined by the authors’ contributions.