A desonestidade científica e seu reflexo na retratação de artigos indexados na base Web of Science, no período de 1945 a 2015

Abstract

The objective of the current study was to analyze retractions of scientific articles in the field of Engineering through articles indexed in the Web of Science database and published between 1945 and 2015. The experimental research consists of two groups: the research group and the control group. The research group is composed of scientific retractions published in the Web of Science database from 1945 to 2015. The target documents were those whose title included the following words: “retracted” or “retraction” or “withdrawal” or “redress”. On the other hand, randomly selected articles constituted the control group; they were selected from the same periodical, number, and volume of the research group articles. The data collected sought to address the study’s goals, such as analyzing the amount of citations and self-citations, the profile of the retracted articles, density, and obsolescence of the publications. To analyze the data, the Kolmogorov Smirnov, Mann-Whitney, and Tukey-Kramer tests were used, along with descriptive statistics. 238 retracted articles and 236 non-retracted articles were analyzed. The analysis shows that the retracted articles have been quoted 2,348 times, with 1,291 citations before retraction and 1,057 after retraction. Furthermore, the retracted articles were quoted 9.87 times, on average. The non-retracted articles received 2,957 citations, with an average of 12.53. In the matter of self-citation, the retracted articles had 555 self-citations, of which 481 occurred before retraction, and 74 after, with an average of 2.33. In contrast, the non-retracted articles received 396 self-citations, with an average of 1.68. Mostly, the articles included an average of 35.5 references and the mean lifespan was five years. The main reasons for retraction are unethical research, in 81 cases, and plagiarism, with 66 cases. Usually, the editors publish the retractions within one or two years after article publication. The study verified that the retracted articles are less cited than non-retracted articles. Essentially, the results suggest that retracted articles receive more self-citations than non-retracted articles, while the number of self-citations is higher before than after the retraction. As such, it is necessary to emphasize that researchers and periodicals may put science’s credibility at risk while preventive and corrective methods are not truly effective in combating scientific misconduct.