Can you trust what you read?

Home Technologist 06 Can you trust what you read?

Thanks to major European initiatives, scientific publishers are feeling the pressure to crack down on plagiarism.


Italian orthopaedic scientist Bernardino Saccomanni is a notorious example of a serial plagiariser. Multiple papers of his – containing entire passages from work published by other authors – reached publication in various journals before being retracted.

It is staggering that such blatant acts of theft slip through the net despite the existence of plagiarism detection software, but bold disregard for the scientific process is far from rare. “A colleague of mine was sent a paper just last week to referee,” recalls Mhairi Stewart, Public Engagement Officer at the University of St Andrews. “It seemed familiar, then she realised it was word for word one of her papers with the formulae changed.”

A recent meta-analysis of surveys put the proportion of scientists who have witnessed plagiarism at around 30 per cent, with 2 per cent actually admitting to it. “I think it’s more that we’re getting better at catching it,” says Ivan Oransky, founder of the Retraction Watch blog. “Also, thanks to the web and open access, there are more eyeballs on papers now that they’re not sitting in dusty libraries.”

More reliable literature

Scientific publishers are feeling the pressure to up their game. Heidelberg-based EMBO, an organisation that supports researchers throughout Europe, has created a Source Data initiative that hopes to render scientific literature more reliable. Meanwhile, the international Reproducibility Initiative – aiming to replicate highly cited cancer biology studies – has taken an average of two months to source data for a paper. The move to request original data as standard would discourage plagiarism. Ultimately journals may need to shift their focus away from publishing what they perceive as highly newsworthy papers towards more complete reporting. Relatively new umbrella organisations such as the UK-based EQUATOR Network are uniting researchers, editors, reviewers, funding bodies and institutions to encourage more reliable reporting. “The biggest challenge is the culture in academia,” says Virginia Barbour, chair of the UK’s Committee on Publication Ethics, “in which all the incentives are to publish as much as possible in the most prestigious journal you can. We know that scientists feel pressured to cut corners, and this contributes to sloppy practices.”

Publish or perish

Yet many institutions continue to drive the “publish or perish” culture so heavily associated with plagiarism (not to mention more serious forms of misconduct). Success, grants, promotion – even cash bonuses – are inextricably linked to a scientist’s publication record. Bernd Pulverer, chief editor of The EMBO Journal, says scientists need much more education on research and publication ethics, calling their level of ignorance often “astounding”. Barbour argues that funders and institutions must address the culture of academia to remove perverse incentives and reward good behaviour. Others have suggested that in-house teams should set standards and monitor publishing integrity. Such approaches would do well to consider educating undergraduates too, to extinguish the temptation to plagiarise before it becomes habitual.

“We’re getting better at catching it. With the web and open access, there are more eyeballs on papers now that they’re not sitting in dusty libraries.”

Institutions should also provide more support for students and researchers not confident in writing about their work. A recent study by information scientists Daniel Citron and Paul Ginsparg claims that those not highly proficient in English are more likely to lift lengthy phrases in order to explain themselves.

Acting unilaterally

EMBO’s Pulverer believes institutions and funders should penalise researchers, establish motives and identify the level of intent. “However,” he says, “institutions do not always pursue independent, formal and informed investigations – on occasion one feels they may have a conflict of interest. While we always try to coordinate with institutional investigations, in these cases we find ourselves having to act unilaterally at the journal level.”

The overwhelming message is that the scientific community needs to tackle plagiarism with a united front, with all researchers having a role to play in detecting and reporting breaches of publication ethics.

Predatory publishers How do ruthlessly plagiarised papers get to publication without detection? In some cases, publishers “relax the rules” for authors who have previously published with them. And at some journals, not all papers are put through the full scope of pre-publication screens.

To date, there’s no overall body that sets standards and processes in the global industry of academic publishing. With low start-up costs, anyone can set up a legitimate-looking journal. “Many operate essentially as vanity presses and will publish almost anything for money,” complains academic librarian Jeffrey Beall, whose blog Scholarly Open Access is filled with such examples.

Powerful software

iThenticate finds pasted sequences of words by comparing manuscripts to online documents and millions of papers in the CrossCheck database.

SciDetect, a collaboration between Springer-Verlag and the Université Joseph Fourier in Grenoble, is open-source software that unmasks manuscripts created with fake-paper generators, such as SCIgen, by analysing word frequency.

Such tools are powerful, as Ed Nutting, a computer science student at the University of Bristol and self-taught programmer, can attest. The software detected similarities between new code he was writing and online work he had done years previously. “In essence, I was accused of plagiarising myself!”

By Holly Cave @HollyACave

References • • the-world • • • • • 100705/full/466167a.html • • • • 2010/100705/full/466167a.html • effort-1.17694 • Often-Due-to-Plagiarism–Study/ • • • • • • •


Angry Birds Finland

The country has a knack with mobile phones. Not only was it home to industry giant Nokia, but it’s now…

Bitcoins have taken one step closer to becoming real money.
Emojis and emotions

With 64 emojis and DKK 1.2 billion tweets, an algorithm has learned to recognize sarcasm in text messages. The method…