Above: Image © istockphoto.com/retrorocket
Did you know? The number of scientific papers retracted in the Web of Science database has risen from approximately 30 a year in 2000 to over 400 in 2011. Cory Toth was a successful physician and medical researcher based at the University of Calgary. In 2013 and 2014, nine of his published scientific articles were retracted because they contained false data. He has since resigned from his position at the university.
Toth’s case raises the question of academic dishonesty, which can be a rather vague concept to people who aren’t university researchers. So I’ll try to paint a clearer picture of what dishonesty in science actually looks like.
Science is all about data. After an experiment, you look at the numbers and attempt to interpret what they are telling you. Data and how they are represented lie at the heart of academic dishonesty, which involves making the results of your research say something they don’t really mean.
Did you know? Some scientists and editors are encouraged by the growing number of scientific papers being retracted. They see it as evidence that the system is becoming better at detecting fraud and that scientists are admitting their mistakes when appropriate.For example, Toth’s paper “Local erythropoietin signalling enhances regeneration in peripheral axons” was retracted because “the author has become aware that Figure 9 had been manipulated.” But how can a figure be manipulated? And why wasn’t the author aware?
In scientific journals like the ones where Toth’s articles appeared—such as Nature, Cell, Science, and the Journal of Molecular Medicine—figures are often images of cells, tissues, and organisms. They can also be graphs or tables that allow readers to visualize the results of experiments.
It’s not clear exactly what was manipulated in the figures published by Toth, but there are a number of ways that a figure can be altered in an academically dishonest way. For example, scientists commonly use software like Photoshop to enhance pictures taken with a microscope to make them easier for the reader to see and prepare the image for publication.
However, there are a number of rules that must be followed. Researchers should only adjust the brightness, contrast, and/or colour balance of the entire image. And they should avoid removing or adding any elements without explicitly stating, in the legend, how the image was altered. If scientists don’t follow the rules, their images will no longer properly reflect the research data, and readers will be provided with fraudulent information.
It is important to note that lead investigators like Toth may not initially be aware of problems with figures, since the data and images are often generated by students under their supervision. However, lead researchers are responsible for supervising anyone working in their labs and ultimately bear responsibility for any mistakes.
Did you know? Only 0.02% of academic papers in the PubMed database are withdrawn because of misconduct. That works out to one out of every 5000 papers. Another way data can be misrepresented is when scientists select which data to include in their analysis so they get the results they are looking for. For example, a researcher might be working with the hypothesis that a certain protein enhances the sprouting of neurons in a cell culture. They would repeat the experiment a number of times, like a good scientist would, and might find that, three out of five times, adding the protein to the culture succeeded in enhancing the sprouting of neurons.
In this case, it would be unethical for the scientist not to report the negative results and only include the cases where the experiment worked. That is, unless there was a good reason for not including the failed experiments, such as evidence that the culture was contaminated or that the cells had died before the protein was added. This is why keeping detailed notes on experiments is very important, so that the exclusion of any data can be properly justified later on.
What motivates a cheater?
Now that you have a better idea of what “academic dishonesty” means, you may be wondering why a scientist would ever compromise their research and risk their reputation in these ways.
Science can be very competitive, with a growing number of qualified applicants looking to fill a limited number of positions. And funding agencies, employers, and journals reward positive results. The scientists whose research yields the most interesting or promising results tend to get the most funding, the best jobs, and published in the most prestigious journals.
A famous case of academic fraud is that of Eric Poehlman, who specialized in research on obesity, menopause, and aging. He ultimately faced criminal prosecution for falsifying data by fabricating data points so his results would appear significant. After denying the accusations for several years, he eventually plead guilty and admitted to presenting fraudulent data in lectures and in published papers. As a result, a number of his papers had to be retracted.
Yet Poehlman had used his research results and reputation to obtain millions of dollars in grants from the US National Institutes of Health. At the time of his prosecution, he was one of the highest paid and most recognized researchers at the University of Vermont. His success highlights just how tempting it must be to manipulate data to make research results more interesting, as well as the extent of his fall from grace.
The cost of dishonesty
Did you know? Software is making it easier to detect plagiarism and image manipulation in scientific papers. This could reduce the number of retractions by identifying fraudulent papers before they are published.Academic dishonesty doesn’t only hurt those cheaters who happen to get caught. It also has nasty consequences for the scientific community as a whole. Granted, the scientific process is designed to be self-correcting: peer review and the ability of scientists to replicate one another’s results are supposed to weed out inaccurate and fraudulent research. But sometimes this takes time and sometimes it doesn’t work. And, in the meantime, fraudulent research can lead other scientists astray.
Scientific progress is achieved brick upon brick. New experiments rely on past ones to generate new hypotheses and to refine methods that have worked in the past. If the foundation is weak, or false, new science can become misguided, wasting time and delaying new discoveries. And public safety can be put at risk when things like medical treatments or safety regulations are based on fraudulent or inaccurate research.
Ultimately, cases of academic dishonesty can serve to shake public confidence and trust in science. Thankfully, in relation to the vast numbers of scientists working around the world and the large number of academic papers papers published every year, academic fraud appears to be a rare phenomenon. But there is always room for improvement. How do you think academic fraud could be better prevented?
An Unwelcome Discovery (2006)
Jeneen Interlandi, The New York Times
Detailed magazine article on the Eric Poehlman case.