Image duplication in published scientific papers

What is it about?

The reproducibility of biomedical science has been a topic of increased concern. Science builds upon science, and the reliance on previous studies can only succeed if published work is of impeccable quality. However, faulty data, whether the result of honest error or intentional misconduct, can result in irreproducible research. Our paper examines a particular form of problematic data – inappropriately duplicated images. We have examined more than 20,000 published research papers published in 40 biomedical journals, screening for simple duplications, duplications with repositioning, and duplications with alterations. Approximately 4% of papers were found to contain some form of image duplication. We not only determined the percentage of problematic images in specific journals, but also examined the influence of factors such as impact factor and country of origin. Among our most important findings is the first empirical evidence that the incidence of problematic data in published articles has in fact been rising over time, with a substantial increase in 2003.

Featured Image

Why is it important?

The scientific literature is cumulative, and the reproducibility of biomedical research is a topic of increasing concern. Inaccurate data, whether resulting from honest error or intentional misconduct, can contribute to research irreproducibility, but the prevalence of such data is unknown. Here, we provide for the first time an estimate of the percentage of a specific type of inaccurate data, inappropriate image duplication, in published biomedical research papers. Approximately 1 out of every 25 papers was found to contain some form of problematic image, with a substantial increase in prevalence since 2003. Current standards appear insufficient to prevent flawed papers from being published. Greater efforts are needed to ensure the reliability and integrity of the research literature.