Collaborating Authors

[Editorial] Research integrity revisited


The U.S. public and private sectors invest billions of dollars and countless hours of highly skilled labor into scientific research every year, an investment that delivers enormous benefits to society. Integrity is indispensable to the orderly and efficient progress of this research. Regrettably, there have been some well-publicized breakdowns in scientific integrity and reported cases of irreproducible research. A new report from the U.S. National Academies of Sciences, Engineering, and Medicine (NASEM), Fostering Integrity in Research, recommends specific steps to secure a future based on integrity and reliability ( These include establishing a new Research Integrity Advisory Board (RIAB) and taking stronger actions to discourage and eliminate practices that are clearly detrimental to research.

[In Depth] U.S. report calls for research integrity board


A report out this week from the National Academies of Sciences, Engineering, and Medicine urges the U.S. research community to up its game in both investigating allegations of research misconduct and promoting ethical behavior--or risk having the government act unilaterally. The report's key recommendation is that universities and scientific societies create, operate, and fund a new, independent, nongovernmental Research Integrity Advisory Board. The board would serve as a clearinghouse to raise awareness of the issues, as an honest broker to mediate disagreements, and as a beacon to help institutions that lack the knowledge or resources to root out bad behavior and foster good behavior. Other entities are already doing these things, it says--federal funding agencies investigate and punish miscreants who misuse taxpayer dollars, universities train scientists, and scientific societies and journals have adopted ethical standards for their authors and members--but none has research integrity as its sole focus nor covers so much territory. And all of these organizations need to do a better job.

Step up for quality research


In response to calls for change from within and outside the scientific community, funding agencies, journals, and professional societies are developing new requirements to promote reproducibility and integrity in research. Amid this activity, the voices of academic institutions--both their leadership and rank-and-file faculty--have largely been quiet. Yet journals, societies, and funding agencies are not in the laboratory, clinic, or field. They do not analyze data, write manuscripts, or prepare figures. Institutions that comprise the global academic community can do more to help their researchers produce the highest-quality results.

Putting my grad school angst to use


My phone rings right when I'm about to leave work, as if the person on the other end has been debating the call all day. On the line is a graduate student. At first they are hesitant to talk, but they loosen up when I assure them their question is reasonable and their dilemma is common. The emotion in their voice makes it clear that just going over university guidelines won't be enough. We talk for about 30 minutes, discussing the details of their research.

Estimating a Null Model of Scientific Image Reuse to Support Research Integrity Investigations Machine Learning

When there is a suspicious figure reuse case in science, research integrity investigators often find it difficult to rebut authors claiming that "it happened by chance". In other words, when there is a "collision" of image features, it is difficult to justify whether it appears rarely or not. In this article, we provide a method to predict the rarity of an image feature by statistically estimating the chance of it randomly occurring across all scientific imagery. Our method is based on high-dimensional density estimation of ORB features using 7+ million images in the PubMed Open Access Subset dataset. We show that this method can lead to meaningful feedback during research integrity investigations by providing a null hypothesis for scientific image reuse and thus a p-value during deliberations. We apply the model to a sample of increasingly complex imagery and confirm that it produces decreasingly smaller p-values as expected. We discuss applications to research integrity investigations as well as future work.