Faked Research Papers Found in Scientific Journals

Faked Research Papers

Faked research papers are being found in academic journals across numerous disciplines, according to the scientific journal, Nature. Not plagiarized. Not based on fraudulent studies or false data. Just plain faked, page after page of gibberish that, when analyzed by a specialized computer program, turn out to have absolutely no meaning at all.

More than 120 of the faked papers have been identified so far, according to Cyril Labbé, a computer scientist at the Joseph Fourier University. Labbé is very well qualified for this kind of detective work.  He created software  that is able to identify faked papers automatically by analyzing the language structure of the articles to test whether an article is real, plagiarized or a complete fake.

The faked research papers Labbé has identified may have been created by a computer program.   The program, SCIgen,  was developed by researchers at the Massachusetts Institute of Technology to show that conferences and journals would accept bogus papers as though they were the real thing. SCIgen works by combining scientific expressions with random phrases and collating these into impressive looking documents.  The program is available online, posted there to encourage others researchers to create bogus papers that would help keep journal and conference editors on their toes looking for more fakes.

So, there it is:  the future has arrived again.  One computerized program is creating bogus scholarly articles, while another one tracks them down. In effect, then, there are computer programs reading and evaluating articles written by other computer programs. No humans need apply.

In 2010, Labbé used the SCIgen program himself to create 102 faked  research papers purportedly written by a fictional scientist named Ike Antkare. (Look at the name carefully.) He added the faked papers to the Google Scholar database, earning Antkare an h-index of 94, making “him” the 21st most highly cited scientist in the world in 2010. Pretty good for a guy who does not exist.

The journal Science conducted a sting operation in 2013 in which Science contributing editor John Bohannan, who holds a real Ph.D in molecular biology, submitted a deliberately faked research article to 305 online journals. An amazing 157 journals accepted the bogus article, despite the fact that the article was purportedly written by authors who did not exist from an African university that did not exist either. The article was translated into French using Google Translate and translated back into English to make it sound like it was written by a non-English speaking author.

Bohannan only submitted his faked research papers to online, open-access journals which, unlike traditional print-media journals, have no limits on the number of articles they can accept and publish. Traditional print-media journals, which are almost always subscriber-supported, have a very active and knowledgeable readership, people who are experts in the field the journal covers, making it virtually impossible for would-be scholars to get away with outright frauds….most of the time.

Labbé and Bohannan have merely scratched the surface of a much more serious problem, because students are becoming professors and professors are getting tenure through this process of “spam publishing” since many academic institutions don’t actually read the papers their candidates claim to have written. They merely count them, and the easiest way to do that is by looking the candidates up on Google Scholar, which does not have the ability to determine whether the periodical, or the article, are bogus or not.  It just lists them all. 

Part of the rationale for why it works this way now is that, in modern academia, specialties have become so abstract that one micro molecular biologist, for example, might not have any knowledge about the area of expertise of another micro molecular biologist so that neither can criticize the other’s work.

SCIgen is just one example of an entire class of software, automatic word generation programs, sometimes called spinner programs, which are readily available on the internet. The target market for this software ranges from academics struggling to cope with the “publish or perish” mentality prevalent in many institutions of higher learning, to “webineurs,” website entrepreneurs who need to generate “original” search engine optimized articles that will improve their search engine rankings and bring more people to their websites.

Taking advantage of the pressure to publish or perish, a whole new industry of online open access journals has sprung up, offering academics the opportunity to publish their works on their websites….for a fee that ranges up to several thousand dollars per item. There are important differences  between these open access “journals” and real, peer-reviewed subscription supported journals. Real printed journals have periodic deadlines, and a very specific amount of space per issue.  Online journals have no deadlines, no page limits and no limit to the number of articles they can accept and therefore have no incentive to discriminate between the worthy and the not-worthy.   No one reads the articles in the open-access journals unless they have a specific interest in the subject of a specific article which is how the operators of these websites are getting away with it.

Ulrichsweb.com, an online directory to scholarly journals of all kinds, put the total number of scholarly periodicals looking for content at 47,845.  Combine a voracious market of online periodicals looking for articles, and fees, with academics struggling to meet publishing requirements, and SCIgen-like  programs  to help them generate publishable material, and you have a prescription for disaster.

Here’s the problem in a nutshell:  as the virtual avalanche of worthless papers and articles floods the market, making it increasingly difficult to identify legitimate research and reliable researchers,researchers are waiting time combing through useless data. Corporations are making hiring decisions based on publishing records. Attorneys may use bogus citations in court pleadings and, most troubling of all, incompetent scholars are getting tenure, all on the basis of  these faked research papers. Something to think about, while sitting in a doctor’s office waiting for an appointment. 

By Alan M. Milner

Scientific American
Nature
Nature World News

3 Responses to "Faked Research Papers Found in Scientific Journals"

  1. Mike   March 3, 2014 at 10:56 pm

    Always had my doubts over the literacy levels of the editors of scientific journals. Now we know they don’t actually bother to read the stuff they publish … Or if they do they can’t tell it’s gibberish.

    Reply
  2. Halldor Fannar Sigurgeirsson   March 3, 2014 at 9:36 pm

    Wow. Talking about corruption in the learned community, huh? Great article, Alan! And an extremely important topic.

    Reply
  3. Mathew Paust   March 3, 2014 at 1:39 pm

    Frightening.

    Reply

Leave a Reply

Your email address will not be published.