The enduring influence of retracted research: determinants and citation patterns across three disciplines

Abstract

Scientific retractions play a crucial role in maintaining the integrity of academic research. However, retracted articles continue to be cited, raising concerns about the persistence of flawed research in scholarly discourse. This study investigates the determinants of post-retraction citations in three disciplines—Business, Social Sciences, and Technology—using a dataset of 531 articles from the Retraction Watch database. We enhance the data with citation metrics from Scopus and Semantic Scholar and institutional prestige from the Academic Ranking of World Universities. Employing a self-controlled case series approach—using each article’s pre-retraction citations as a baseline—we apply negative binomial regression for rate modeling to compare citation patterns before and after formal retraction. Our analysis reveals that retracted papers published in prestigious journals and authored by scholars from top-ranked institutions receive more post-retraction citations. Moreover, leveraging Semantic Scholar’s citation types (Highly Influential, Background, Methodology, and Results), we assess how retracted articles continue to shape subsequent research. We found that the prestige of journals and authors’ institutions also influence highly influential and background citations. Additionally, the severity of misconduct is not significant in any of the models, implying that the scientific community in these three fields appears unaware of questionable research practices. These findings underscore the need for better mechanisms to prevent the spread of erroneous research and raise important questions about the role of academic prestige in the continuation of flawed science.