Image manipulation is an ongoing concern in scholarly publishing. Since its launch in 2010, Retraction Watch, a blog that reports retracted scholarly papers, has reported hundreds of papers retracted from the literature because of image concerns. Further, during an Elsevier Researcher Academy webinar on the topic, 52% of the audience rated image manipulation as a “very serious” challenge, 36% rated it as “somewhat serious,” and only 18% didn’t see image manipulation as a serious problem.
Some authors manipulate images for clarity and without malicious intent. According to this Elsevier post, some minor adjustments or enhancements are acceptable, such as simple magnification or additions of arrows and highlights without data distortion. However, some authors manipulate images to complement a desired outcome, an intentional act of misconduct. Unacceptable examples are described in this article from the Journal of Cell Biology and include “cleaning up” an image by erasing backgrounds, editing specific features for selected enhancement, inappropriately adjusting the resolution of an image, and not reporting modifications in figure legends.
Elsevier suggests that an automated system is the best solution for catching instances of image manipulation and urgers stakeholders to prioritize programming one; available tools are at experimental stages and not yet user friendly. Automation can add a level of scrutiny beyond sole reliance on the eyes of journal editors and peer reviewers, but until such technology launches, resources are available to help spot problem images. Seen here, here, and here, Retraction Watch started a “Forensic Fridays” series that provides practice in recognizing questionable images.
Best practice? Always submit accurate, appropriate, and unmanipulated graphics.