Research shows dynamics of cross-platform attention to retracted papers

Evanston [US], June 15 : A study of the online attention received over time by 3,851 retracted scientific articles found that retracted articles received more attention on social media, news outlets, and knowledge repositories, including more critical attention on Twitter, compared with non-retracted articles.
According to the study by Northwestern University, the reason can be attributed to the fact that most articles were retracted after the attention they received had diminished, the results suggest that retractions alone may not be effective at curbing the spread of misinformation online, according to the authors.
Retraction in academic publishing is an important and necessary mechanism for science to self-correct. Prior studies have shown that the number of retractions has increased in recent years. This rise can be explained by many different factors. A reason is that the number of publications is increasing exponentially.
Meanwhile, as scientific research has become more complex and interdisciplinary than ever before, reviewers are facing a higher cognitive burden. This undermines the scientific community's ability to filter out problematic papers. In fact, research shows that prominent journals with rigorous screening and high publishing standards are as likely to publish erroneous papers as less prominent journals. Finally, not all retractions are due to research fraud--some papers are retracted due to unintentional errors or mistakes, which become more likely as research data grow in size and complexity.
Regardless of the reasons behind this increase, a high incidence of retractions in academic literature has the potential to undermine the credibility of scientific communities and reduce public trust in science. What is more, the circulation of misleading findings can be harmful to the lay public, especially given how broadly papers can be disseminated via social media.
For instance, there are two retracted papers among the ten most highly shared papers in 2020 according to Altmetric, a service that tracks the online dissemination of scientific articles. One of them, published in a top biology journal, reported that treatment with chloroquine had no benefit in COVID-19 patients based on data that were likely fabricated. Another paper, published in a well-regarded general interest journal, falsely claimed that having more female mentors was negatively correlated with the post-mentorship impact of junior scholars. Both papers attracted considerable attention before they were retracted, raising questions about their possible negative impact on online audiences' trust in science.
As these examples suggest, retracted papers can attain substantial online attention, and potentially flawed knowledge can reach the public, which often is impacted by the research results. This large-scale spreading of papers occurs as the web has become the primary channel through which the lay public interacts with scientific information. Past research on the online diffusion of science has mainly studied the spread of papers without regard to their retraction status. Other work has examined the dissemination of retracted papers in scientific communities, focusing mainly on the associated citation penalty.
However, the impact of retraction on the online dissemination of retracted papers is unclear. Here, we address this essential open question. Past research found that authors tend to keep citing retracted papers long after they have been red-flagged, although at a lower rate. This raises the question of whether the retraction is effective in reducing public attention beyond the academic literature. Studying the impact of retraction relative to the temporal "trajectory" of mentions a paper receives could be helpful for journals to devise policies and practices that maximize the effect of retractions.
To understand whether the retraction is appropriate for reducing dissemination online, we first assess the extent of online circulation of erroneous findings by investigating variations in how often retracted papers are mentioned on different types of platforms before and after retraction.
Recent research indicates that, overall, retracted papers tend to receive more attention than non-retracted ones. Prior work also showed that retractions occur most frequently among highly cited articles published in high-impact journals, suggesting a counterintuitive link between rigorous screening and retraction.
Is there a similar tendency online where retracted papers receive more attention on carefully curated platforms, such as news outlets, than on platforms with limited entry barriers, like social media sites? Such a trend would highlight difficulties with identifying unreliable research given their broad visibility in established venues and could inform attempts to manage the harm caused by retractions. Second, we distinguish between critical and uncritical attention to papers to uncover how retracted research is mentioned. More than half of retracted papers are flagged because of scientific misconduct, such as fabrication, falsification, and plagiarism. These papers may receive lots of attention due to criticism raised by online audiences.
Is the attention received by retracted papers due to sharing without knowing about the mistakes of a paper, or is it rather expressing concerns (so-called "critical" mentions)? As suggested in a recent case study, knowing how retracted papers are mentioned may uncover users who are improving science-related discussions on Twitter by identifying papers that require a closer examination.