How many research frauds are women?
Origins of this post
In January 2014, two papers were published in the top science journal Nature which claimed that stem cells could be created by simply exposing ordinary skin cells to mild acid shock Obokata et al 2014.The first author on both of these papers was a young Japanese stem cell scientist, Haruko Obokata, working at the Riken Institute in Tokyo. I skimmed BBC news reports of these findings because I realised that this would make virtually unlimited supplies of stem cells available for research and stem cells of a patient’s own tissue type (i.e. not susceptible to rejection) readily available for therapeutic uses. The response of senior researchers quoted in this BBC news report was a mixture of excitement at the prospects and shock at the apparent ease and simplicity of the process. However, within a few weeks of their publication doubts began to emerge about the veracity of the findings and on 1st April 2014 the results of a formal investigation by the Riken Institute were released stating that Obokata had used fraudulent actions in the delivery and publication of these astonishing results. These papers were formally retracted from Nature when it became clear that the results could not be replicated by other scientists at the Riken.
By July 2014 when these papers were retracted, I had researched about two dozen important cases of research fraud as part of an ongoing project to produce a book about error and fraud in the biological and medical sciences. I added this case to my list of future case studies and noted that this was the first time that a female scientist had appeared on my list of shame. Even now there are only 3 women on my list and all of them have been accused since 2005:
- Kristin Roovers at the University of Pennsylvania in 2005
- Haruko Obokata from Tokyo in 2014
- Anna Ahimastos from Melbourne in 2015.
My cases were not identified in any methodical way, they were just cases that I had come across and arbitrarily decided to follow-up because of their importance or because of their interest to me. I therefore looked for independent corroboration of my strong impression that research fraud is or has been largely perpetrated by men.
Results and conclusions from my investigations
In the appendix to their classic 1982 book on research fraud (The Great Betrayal. Fraud and Deceit in Science) William Broad and Nicholas Wade produce a summary of “known or suspected cases of scientific fraud” dating from the 2nd century BC right up to 1981. In their list of 37 names almost all (34) are definitely male and the three where gender is not clear are almost certainly also male. Despite the limitations of our selection procedures, both myself and Broad and Wade have found that in historical cases of research fraud, men have almost always been the perpetrators. Broad and Wade’s list is now 35 years old and it is not surprising that men dominate because until the last few decades there were relatively few women scientists and very few in senior research roles. To illustrate this point, if one looks at a list of those honored by being elected Fellows of the Royal Society in London then just over 150 of the 8000 or so people elected since the 17th century have been women. The first women were not elected until after World War 2 (Kathleen Lonsdale and Marjory Stephenson in 1945). Of the total of 154 women elected to date, 90 (59%) have been elected since 2000.
I found two more recent sources that give lists of research fraud perpetrators:
- Ivan Oransky in June 2015 published an unofficial Retraction Watch Leaderboard of the 30 scientists with the highest number of retracted papers. Only two women appeared on this list in 24th place (Marion A Brach) and 26th place (Silvia Bulfone-Paus) with respectively 14 and 13 retracted papers. This list has recently been updated and now seems to have only one woman on it, Erin Potts-Kant in 28th place (16 retractions). The two women listed on the earlier version have been displaced by other people with more retractions. Note that top of this leaderboard is Yoshitaka Fujii with 183 retractions, Joachim Boldt is in second place with 94 and all of the top ten have more than 30 retractions each.
- Wolfgang Straube and two Dutch psychology colleagues (2012) were interested in how research fraud was discovered and how fraudsters were unmasked. In order to do this, they identified what they referred to as 40 “notorious cases of scientific fraud” dating from 1974 to 2012 where the case had attracted enough media attention for the mode of discovery to be readily identifiable e.g. from newspaper reports. Ten of these forty cases involved a female perpetrator.
Number of retracted papers is an objective measure but it does not always fully reflect the magnitude or seriousness of the offences committed by a fraudster. In fact multiple paper retractions was relatively uncommon in the past even for people clearly acknowledged as major fraudsters. Only 10 of the 43 people on Stroebe’s list of notorious cases make it onto the Retraction Watch Leaderboard and many on my personal list of fraud case studies are also missing from this Leaderboard including:
- Ranjit Chandra, a Canadian physician and erstwhile authority on nutrition and immunity has only one retracted paper but is widely believed to have been fabricating data for up to two decades and was the subject of major investigation by the Canadian Broadcasting Corporation .
- Werner Bezwoda a South African cancer specialist who fabricated clinical trial data which falsely suggested that a very aggressive, dangerous, distressing and expensive high dose chemotherapy regimen greatly survival time in women with advanced breast cancer. Only one of his papers has been formally retracted.
- Ram Bahadur Singh is an Indian physician and nutritionist who has published hundreds of scientific papers many of which are regarded as suspect yet none have been formally retracted. His papers have resulted in major investigations and “statements of concern” by two of the world’s leading medical journals (Lancet and British Medical Journal) and data submitted to support one of his rejected papers was shown to have been fabricated.
Despite the limitations of the measures and methods, these sources allow me to confidently conclude that:
- There are very few women among the historical cases of research fraud.
- Women are infrequently found among what might be called the “career fraudsters” who have built successful careers around the routine fabrication of data. Some of these career fraudsters have published fabricated or probably fabricated data over decades and in some more recent cases have had dozens of their papers retracted.
- Women have infrequently been the perpetrators of notorious cases of scientific fraud that have attracted the intense attention of the popular media.
Women are now much more common in the ranks of professional scientific and medical researchers and substantial numbers are now reaching senior research positions. Are today’s large crop of female researchers less likely to commit research fraud than their male counterparts? Ferric Fang and two colleagues (one female) at the University of Washington in Seattle (Fang et al, 2013) addressed this question . They looked at over 200 individuals who had been identified from records of the Office of Research Integrity in the USA as having committed research fraud. They found that the guilty were spread across all levels of the research hierarchy but that two thirds of all the fraudsters were men. They compared the sex ratio of those convicted with the proportions of people at different levels of the career hierarchy and different categories of science. They found that in both life sciences and in the physical and engineering sciences and at all levels of seniority that many more men were present than would have been predicted from a simple gender breakdown if both sexes were equally likely to commit fraud. This gender disparity was particularly prominent among those classified as senior scientists. They go on to discuss possible reasons for this gender disparity and highlight the observations that men are more likely to engage in risky behaviours generally and also more likely to commit crimes than women.
In a commentary upon this article by Fang and his colleagues Anna Kaatz and two colleagues (one male) from the University of Wisconsin (Kaatz et al , 2013) suggest that Fang’s conclusion that men are more likely than women to commit more fraud is not totally secure. Fang et al (2013) in a rather throwaway line do suggest that it is possible that women may commit as many offences as men but are less likely to be caught and censured which Kaatz and her colleagues suggest is a real possibility. They argue that gender stereotypes may make investigators less willing to accept that women have committed fraud and thus require a higher standard of evidence before finding the accused guilty. They also suggest that that as women are generally considered more likeable than men and more likely to apologise for any wrongdoing, then some may escape public censure by apologising and thus persuading investigators to halt the process before it has reached the stage of a formal Office of Research Integrity investigation and report. They point out that there is also gender disparity in receipt of NIH (US government) funding for research which means that men may have more opportunity than women to commit serious research fraud. It is refreshing example of gender fairness to see two women based in a Center for Women’s Health Research use gender bias arguments and inequality of opportunity arguments to defend the integrity of male scientists!
Despite these reservations it seems probable to me that, as in the past, women are still much less likely to commit research fraud than their male counterparts.
My next post will be a case-study of a research fraud committed by a woman.