Drawing on insights from a recent international survey on research integrity and a recent high-profile case, Nick Allum and Robin Brooker find previous work on scientific plagiarism may have underestimated its prevalence.
The resignation of Claudine Gay as President of Harvard University has pitched the subject of research integrity into the mainstream. Gay, along with the presidents of MIT and UPenn, fell foul of questioning from Rep. Elise Stefanik (R-N.Y.) at a House committee hearing on antisemitism in November 2023 concerning their institutions’ response to calls for the genocide of Jews. However, if demands for her resignation had simply concerned her performance in front of the House committee, there is an even chance she would still be in post. It was the additional questions surrounding her scholarship and the opening salvo of what has been called the ‘plagiarism wars’ that tipped the scales in favour of her departure.
Plagiarism can be defined as:
“When somebody presents the work of others (data, words or theories) as if they were his/her own and without proper acknowledgment”.
Claudine Gay’s work in political science is currently facing rigorous examination for possible plagiarism. Recent reports have revealed as many as 50 instances of plagiarism in her work, stretching back to the 1990s and encompassing her PhD thesis.
Appropriately acknowledging the work of others is fundamental to research integrity in a well-functioning scholarly system but behind this broad definition there are numerous forms of plagiarism, some of which are arguably more or less damaging. For example, one of us (Nick Allum) discovered recently that a co-authored article published in a scholarly journal was completely replicated and published under different names in another journal. A particularly egregious example, encouraged by the rise of predatory publishers with little regard for research integrity.
In Gay’s case, many examples involve exact copying of short sections of text without attribution. While this still sounds like an open and shut case, there have been variable responses as to its seriousness – not least from researchers, such as David Canon, whose work Gay now admits to copying. Some of the examples could be regarded as minor infractions as they are technical descriptions of statistical models, or set out only background information. Arguably, they do not claim an important original insight that is actually due to someone else. Not everyone agrees. For example, Carol Swain, who alleges that her work was demonstrably plagiarised by Gay.
In light of this furore, it is useful to ask: how widespread are the kind of behaviours of which Gay has been accused and in some cases admitted? A recent meta-analysis (combining the results of multiple previous studies) estimated that 2.9 percent of researchers had admitted to plagiarism or other types of misconduct – falsification or fabrication of data. Another study estimates that 1.7 percent of researchers have plagiarised while 30 percent have witnessed others plagiarising. This discrepancy might be due to individuals’ reluctance to admit their own misconduct and the greater likelihood of observing others’ behaviour.
In our recent International Research Integrity Survey (IRIS), part of a larger research integrity project, we asked over 35,000 respondents, amongst other things, about how often they had engaged in questionable research practices (The supporting data can be found on here). One of the specific questions we asked was:
“Thinking about research carried out for your publications over the last three years, how often has the following occurred?
“Deliberately using another researcher’s unpublished idea without giving credit. For example, publishing an idea voiced by a colleague at an informal meeting without giving them credit.”
The response options were often, sometimes, rarely, never. We compared those who answered ‘never’ with all other responses. This is because even someone who reports rarely doing this is still saying that they have done it at least once in the past three years, which, considering academics are unlikely to publish large numbers of papers and books over this period, is not trivial. The question does not ask about copying published work, but captures the essence of plagiarism in that it concerns passing off an idea as one’s own that was actually originated by somebody else.
We found that around 7 percent of researchers across the whole sample owned up to stealing someone else’s idea in this way. This is higher than previous estimates. Perhaps it is because stealing an idea rather than text is much harder to prove and therefore less likely to be detected and punished. This average also conceals some geographical differences. The heatmap below illustrates the varying percentages of researchers in different countries who admit to stealing ideas.
While many countries are in the 4-8 percent range, a few have even higher rates – typically, but not limited to, states that have joined the European Union more recently. Beyond the country where they are employed, it is also interesting to look at what kind of researchers are more likely to report stealing an idea.
In the plot below, we show how much more probable it is for a researcher to report this behaviour based on different attributes. The dots show the average increase or decrease in probability of reporting stealing an idea for someone with that attribute. The bars represent the margin of error (95 percent confidence interval). We model all these attributes together to predict who is most likely to report stealing an idea.
The biggest contrast is between industry and academia, where for those in industry the probability increases by about 3 percent. Non-profit and government researchers are also a little more likely to admit to stealing an idea, as are those working in humanities and social sciences. Of the characteristics associated with a lower probability, being female and expressing stronger adherence to classical ‘Mertonian norms’ of science are the most pertinent, along with working in a more collegial and supportive working environment.
Does Claudine Gay fit this profile? A ‘risk factor’ is that she works in social science (political science). However, she also works in academia – generally the least probable sector where plagiarism occurs. Women also have a lower probability of admitting to taking ideas from others. Unusually, Gay has worked almost exclusively in high status universities (Stanford and Harvard), where the pressure to publish is strong. It is possible that this environment is less collegial and incentives to cut corners to publish may be greater. For whatever reason, she does not fit the profile well.
That said, our models are not very precise in their prediction. Rather, our research indicates that the tendency to engage in questionable research practices is more an individual habit, with minimal correlation to the country, organisation, or observable attributes of the researchers. One thing we do know is that it is relatively uncommon, which makes it a serious accusation that shouldn’t be shrugged off.
Ultimately, plagiarism undermines research integrity. And despite varied views on the severity of Gay’s acts, citing sources is fundamental in scholarship. Her downfall shows that this ethical obligation extends even to those at the top.
The content generated on this blog is for information purposes only. This Article gives the views and opinions of the authors and does not reflect the views and opinions of the Impact of Social Science blog (the blog), nor of the London School of Economics and Political Science. Please review our comments policy if you have any concerns on posting a comment below.