Self-plagiarism: A retrospective study of its prevalence and patterns across scientific disciplines

This study explores the incidence and nature of self-plagiarism across various disciplines. Using Retraction Watch data, we analyze retraction times, authorship patterns, disciplinary differences, and geographical origin of manuscripts retracted due to duplications of data, image, complete article or parts of text. We find that image and entire article duplications are primary causes of self-plagiarism-related retractions. Duplication of entire article is most prevalent in the social sciences and business. In biomedical fields, image duplication is the main issue, with the longest median retraction time of 4 years. In comparison, the median retraction time for all self-plagiarism cases is 3.2 years, and for all plagiarism-related retractions - 1.7 years. We find more authors on self-plagiarized than plagiarized manuscripts. Author counts are also higher when duplicated material is data or image than when duplication involves text or entire article. Our comprehensive analysis not only highlights critical issues but also offers insights for developing stronger policies to uphold academic integrity.