Monday, July 12, 2010

To catch a plagiarist

Scientific dishonesty is all the rage these days. No, the cool kids aren't doing it.

Amidst the public gnashing of teeth at the IPCC over the Climategate report and the science blogosphere reeling from Pepsigate, Nature ran a piece last week detailing how it and other publications had started cracking down on plagiarism in their journals.

The topic isn't new in science, but the ongoing plagiaristgate (copycatgate? cheatergate? Unattributed-recycling-of-previously-used-materials-gate? Sorry, part of being a journalist is unnecessary attachments of the word gate to any noun involved in a scandal.) shows that the problem isn't only confined to high school English classes.

In 2008, several thousand publishers organized a non-profit plagiarism checker called CrossCheck that draws on the collective power of 25.5 million articles from nearly 50,000 journals to catch would-be cheats. To date, 83 publishers have joined the database, with many supposedly holding out over fear of sharing content.

According to the Nature article, one publisher responsible for thousands of publications on a broad range of topics, Taylor and Francis, found that since they started using CrossCheck six-months ago on three of their journals; one had to reject 6-percent of submissions because of plagiarism, another 10-percent, and the third had to reject a whopping 23-percent (13 of 56 submissions). The instances ranged from taking a few paragraphs to just copying and pasting other people's articles entirely. However, the biggest problem they found was self-plagiarism.

The very small number of studies released so far have found much less rampant plagiarism than Taylor and Francis, but the most frequent violations are consistently self-plagiarism.

In the past, revealing a scientific fraud has left people sounding the trumpets of science's self checking nature. You can lie, cheat and forge, and you'll still get published. In the end though, someone will check your work. In an era of mega public-funding for scientific research, science can hardly afford to tolerate even these minor injustices when they threaten to undermine public support. The IPCC researchers were cleared of scientific fraud by several investigations, yet public confidence in all climate change research has dropped. That's despite mounting evidence that shows climate change is a major concern and some in the news media taking the rare step to admit they blew it.

In 2002, arguably the world's brightest young scientist, a physicist named Jan Hendrik Schön, was proven to be a monumental fraud. He deceived the most prestigious journals in the world like Science, Nature and Physical Review Letters (a publication of the American Physical Society who supplies this intern's paycheck), with articles detailing fascinating advances he had made in transistors, lasers and superconductors. His peers did eventually find him out when other researchers had trouble replicating his results and in addition to making things up, he was also found guilty of taking data from one paper and simply dropping it into another (Eugenie Reich has an awesome book out about it).

A Nature editorial also released last week lists some times when it claims plagiarism might be justified and cites human judgment as the ultimate decider on what is and isn't plagiarism. The editorial says sometimes a self-plagiarizing author might be bringing old material to a new audience, or other times a scientist with a poor mastery of English might paraphrase from the introduction of similar work.

Schön's example is far more alarming than a tenure-thirsty professor recycling old material to get their name in print more often. However, public trust for science has been eroded by scandals in recent years and journals at least using CrossCheck are taking steps to make it easier for reviewers to check suspect material.

Most publications haven't released data on how many submissions they're rejecting after checking in CrossCheck. Nature says it's only spot-checking its research articles and finds very little plagiarism, though it checks all of its review articles and has found higher instances there (still less than one-percent).

Someday CrossCheck's use might work as a deterrent for would be cheats, but considering the media's kneejerk reaction to the Climategate story, damage is easily inflicted and incredibly difficult to repair. Every publication should take at least the same steps as a tenth grade English teacher. Every field doesn't need its own cautionary tale of Piltdown man, Climategate or Jan Hendrik Schön.

9 comments:

  1. Flash - CrossCheck is more a user group than a product. iThenticate, the software that powers CrossCheck, is available for licensing by government agencies, corporations, research facilities, authors and individual researchers. In fact, the growth of STM journal publishers using iThenticate has prompted us to introduce an individual license option that will allow researchers to vet single manuscripts via iThenticate before submission to journals for publication.
    -Bob Creutz, iThenticate GM

    ReplyDelete
  2. Thanks for commenting, Bob.

    One thing I'm curious about in Nature's piece is why CrossCheck was formed by thousands of publications, yet only 83 have joined the database. Is it a matter of uploading so many articles or is it truly fears over content sharing?

    This seems like something that could have happened years ago.

    ReplyDelete
  3. Falsh your statements in support of AGW discount the balance of your article as coming from another useful idiot who is either lacking in critical thinking skills and is just subject matter ignorant and lazy in accepting the religion of carbon.

    People have lost "faith" in science because AGW is a massive fraud, and individuals like yourself are too lazy or biased to consider the wealth of evidence which invalidates the hypothesis by Mann, Jones, Hansen, etc.

    ReplyDelete
  4. Anonymous, I'm always amazed at the fraud claims in the hyperbolic arguments against anthropogenic global warming. Fraud implies a motive. I have a hard time seeing what would motivate scientists to lie to support claims of global warming. There's a lot more money (i.e. incentive) to fraudulently claim there is no AGW than the other way around.

    ReplyDelete
  5. Hi Flash - CrossCheck is a joint initiative run by iThenticate and CrossRef. CrossRef is a non-profit organization set up ten years ago to enable reference linking between publishers, but now providing additional services such as CrossCheck. CrossRef has 3000 publisher members, 83 of whom have joined CrossCheck to date. These 83 publishers have added all of their content to the database, and use the iThenticate software to screen their submissions against this database.

    I'm not sure it's true to say that many publishers are holding out over fear of content sharing; I think most would agree the benefits outweigh the very minimal risks. The project is relatively young and still gathering momentum, and publishers are working out how best to integrate plagiarism screening into their editorial processes. We're seeing more CrossRef members join CrossCheck every month, and its effectiveness grows as each new publisher joins and adds content.

    -Kirsty Meddings, CrossRef.

    ReplyDelete
  6. Thank you as well for the comments, Kirsty.

    I'm still not seeing where the negatives come in to keep the other CrossRef journals from joining CrossCheck. You mention that you don't think publishers would hold out over content sharing fears as implied by Nature news, but that leaves the question out there as to why everyone isn't using it. Is it prohibitively expensive for some journals? Nature also suggests that the 75-cent per article fee is considered reasonable by most.

    Two years seems like more than enough time to incorporate plagiarism checking into the editorial process, surely Taylor and Francis demonstrate that it's a good idea.

    If the negatives are minimal and benefits are high, why are only a couple percent of CrossRef members on board?

    ReplyDelete
  7. Something like CrossCheck is sadly needed, but I would be a bit suspicious about calling it "not for profit".

    iThenticate is a commercial company making loads of money out of plagiarism checking, and I wouldn't be surprised if content shared via CrossCheck is then available to the company when scanning for problems in student work. Adding any content to their database just makes their service more effective and their monopoly stronger.

    ReplyDelete
  8. It's worth noting that the 83 publishers in CrossCheck together publish over 60% of the content of all CrossRef publishers, so in terms of number of journals, well over half are participating. Many of the remaining members are very small organisations that publish one or two journals and don't have the resources of Taylor & Francis or Nature.

    You can see a list of participating publishers and download a list of the titles covered in the database on our website: http://www.crossref.org/crosscheck.html

    ReplyDelete
  9. i am regular visitor of your site and you are writing very nice so keep posting university admission

    ReplyDelete