[This post was created for the Wolters-Kluwer author newsletter Author Resource Review and has been reproduced with permission. The content in the post is co-authored by Phil Daly and Ginny Pittman. Phil Daly is a Senior Publisher at Wolters Kluwer Health and has over 25 years’ practice in STM editorial, managing editor and publisher roles. Phil holds […]
The number of published papers being retracted is enhancing dramatically and is higher than ever before. Albeit retracted papers still represent a miniscule proportion of the total published literature, the rate of increase in retractions is alarming, as retracted papers tarnish both a researcher’s career and the integrity of the literature itself.
Studies may be retracted for a diversity of reasons. The reasons for retraction can broadly be categorized as fair error, intentional or unintentional misconduct, or others. The very first two categories and especially interesting, and I’ll discuss them in detail. The third category—others—usually involves papers where the reason for retraction is not clearly stated in the retraction notice or is not clear to the journal editor(s). For example, authors may retract a paper that has been at the same time submitted to and accepted by another journal. Albeit this would amount to misconduct on the authors’ part for not having informed the journal of the duplicate submissions, the retraction would be listed in the “other” category if the authors do not disclose the reason for retraction.
The table below lists possible reasons that may feature under the former two categories—honest error and unintentional or intentional misconduct. These are some of the common reasons, albeit the list is not exhaustive. Most retractions with these underlying reasons are initiated by the journal.
- Errors in sample or data
- Undisclosed conflicts of interest
- Skewed statistical analysis
- Plagiarism or self-plagiarism
- Inaccuracies or unverifiable information
- Salami slicing (using the same data set to publish numerous studies)
- Data fabrication or manipulation
- Redundant publication (discovery that some aspects have already been published)
- Lack of adherence to ethical protocols
- Disputes over authorship attribution
- Duplicate submissions (to different journals at the same time)
Another interesting but alarming fact is the way in which the reasons for retraction have switched over the last thirty years. A paper published in the Journal of Medical Ethics in 2011 reported that 73.5% of 742 retractions inbetween 2000 and 2010 were due to error and only 26.6% due to fraud. However, a large-scale follow-up probe published in PNAS very recently, which involved the author of the 2011 examine as well, found that misconduct (including data manipulation and plagiarism) is actually the primary reason for retraction, accounting for 67% of the total retractions since 1977. The figure below shows the major reasons for retraction and how they have become increasingly prominent since 1977.
(Figure credit: Fang et al. PNAS 2012)
So as you can see, retractions, especially those due to misconduct, are enlargening at an alarming rate. What’s more is that plagiarism and duplicate publication, which were not factors earlier, are becoming more commonplace, possibly because of fresh technologies such as plagiarism detection software packages now employed by journals and because of the increase in submissions from countries where English is not the very first language.
This increase in retractions has been attributed to the pressure to publish and the rat race that researchers find themselves in. With latest studies having shed light on how misconduct is plaguing science, journal editors and peer reviewers are more likely to be alert in spotting issues with papers before they are published.