Why research integrity matters to all of us

Research integrity is a slippery, tricky, and nervously approached topic.

This is partially because the validity of research has a direct impact on the world. It is university research that forms the basis for curing cancers, building atomic bombs, and tens of thousands of other kinds of big decisions that fundamentally shape the world around us. If the research is wrong then, well, then the world might just be wrong as well.

It wasn’t me

The difference between bad practice and outright cheating or fraud isn’t always clear. There is the obvious like copying someone else’s work, or making up results, or otherwise fabricating, lying, or using technology to produce bullshit citations, in the hope of deceiving the wider academic community. However, there are the more subtle practices that veer on the morally suspect and academically poor but harder to call cheating. Quietly omitting unfavourable results, poor citation, over egging the outcomes of some work for the sake of novelty, and other academic malpractices.

Plagiarism of research is also a political issue as starkly shown by the resignation of former Harvard President Claudine Gay.

The Washington Post has a full timeline of the Gay affair but in short following comments made at a congressional hearing on antisemitism on campus which Gay herself later apologised for there was increased attention brought to her work. The Washington Post characterised the accusations as

The reports published in various outlets collect nearly 50 instances in which Gay allegedly misused academic sources. They appear in eight of her works: a 1993 essay in the magazine Origins, her dissertation from 1997, a 2001 working paper, and five articles she published while a professor at Stanford and Harvard (out of a total of 11 journal publications across her career).

In the end, Harvard found that Gay had not committed research misconduct and instead had a few instances of poor citation attribution. Hardly the crime of the century (albeit Carol Swain whose works was allegedly plagiarised may disagree) and certainly a result not coterminous with the amount of attention the case received. Regardless of the case itself it was the accusation of plagiarism that was wielded as the tool to pressure Gay to resign from her role.

Unfaithful

Gay’s case is not the first plagiarism accusation aimed at a university leader. The former President of the University of South Carolina resigned after plagiarising part of a graduation speech. Once Stanford President Marc Tessier-Lavigne resigned after serious issues were found in his research.

Perhaps most famously and damaging is the case of Andrew Wakefield who in a piece for the Lancet alleged there was a link between MMR vaccines and autism. In turn this led to declining use of the vaccine. As it transpired not only were his results impossible to reproduce. Not only did he have undisclosed financial interests in the results. Not only did Wakefield fail to collect results ethically. He was also guilty of “deliberate fraud” in his use of facts and data.

The yawning chasm of practice between Wakefield and Gay shows that one of the challenges with research plagiarism is that it covers everything from the sloppy to the downright dangerous. In her Book My Word! Plagiarism and College Culture Susan Blum writes that plagiarism is not “one thing” but a site of profound disagreement on how the collation, reproduction, and sharing of knowledge should be managed.

Little lies

The prevalence of academic fraud is another point of debate. An international survey carried out by LSE academics suggests that up to seven per cent of academics in the sample committed some form of plagiarism in the past three years (much higher than estimates from previous meta-analysis). A series of freedom of information requests by the BBC found 300 reports of research misconduct at 23 of the 24 Russell Group universities between 2012-15 with around ⅓ of these upheld. The House of Commons Science, Innovation, and Technology Committee’s enquiry into research reproducibility and research integrity found that there are significant reproducibility issues with research but stopped short of calling it a crisis.

Why researchers engage in questionable practice is less explored. Some of the high profile cases emerge from poor scholarship, accidents, or slack checking of sources. Not really the kind of activity that points toward a full blown crisis. There is the kind of hybrid wishful thinking presenting statistics as significant known as p-hacking which serves to show research as impactful. P-hacking speaks to the difficulty of the volume of research that ultimately leads nowhere. And there is the range of statistical manipulation techniques to raise the profile of research for the benefits of personal prestige and promotion.

I’m not the only one

It is impossible to know the scale of the challenges with research integrity but what emerges are kinds of vanity cheating. The why won’t anyone see what I see as important kind of boosterism of work. There is a strain of not checking, attributing and citing kind of research integrity which depending on its nature falls between honest mistake and outright deception. There is a field of having worked hard on something and it’s hard to believe the results don’t work. And, with the likes of Wakefield, there is the hopefully rare outright fraud, lies, and blatant malpractice.

In the end, it is not just the scale or severity of intentional or accidental deception nor its nature that will harm the wider reputation of UK research but the perception that this could be happening. This is only made more complex in a world with AI which blurs the line between gathering ideas and fabricating work. As Dorothy Bishop points out it only takes a few foundational papers to be wrong to undermine our hard won reputation for research rigour.