Tuesday, March 29, 2016

A Confusing Pakistani Plagiarism Case

The Pakistani Express Tribune reported on March 28, 2016 that the former chair of the Pakistani HEC (Higher Education Council) in Pakistan has tendered his apology for having plagiarized a paper he co-authored with a colleague. According to the Express Tribune, an investigation committee found in 2014 that a third of the paper was taken from a European Union (EU) report.

The paper, given as ‘Study of Pakistan’s Election System – Intelligent E-Election’ in the Journal of Independent Studies and Research (Vol. 1, No 2, July 2003, pp 2 – 7), will be "withdrawn" from the author's CV, according to the Express Tribune.

There are two journals that are "JISR" journals, one on management and one on computer science. According to the online table of contents of the management issue, although seven papers were accepted, only four are linked to online. One is indeed by the former HEC chair, but on a different topic. The computing issue (Vol. 1, No 2, July 2003) includes a paper by his co-author on a different topic.  I checked the management journal on the Internet Archive and found the July 2003 issue archived in 2012 with no mention of the paper, but it can be found linked from the table of contents of the computing issue, also archived in 2012. So the paper was removed from the journal instead of being retracted.

If I google the title, however, I land at an arXiv paper by that same name and by those authors, submitted on 27 May 2004, with no mention that the paper has been published elsewhere or that it has been withdrawn. The acknowledgment of the paper is interesting:
By the grace of Allah, this independent study stands complete. However, calling it independent is not quite correct. I owe the success of this effort to my institute and my Independent study supervisor, Dr. J[...] R. L[...], without whose guidance this accomplishment would have been impossible.
So is this a colleague or a student who is the first author of the paper?

Digging a bit deeper turns up this 2013 report in Dawn stating that using Turnitin on the paper reported 78 % plagiarism. I have grave reservations about using numbers that a plagiarism detection system reports, as they are often skewed. The article notes:
According to the HEC rules, anyone found involved in plagiarism before 2007 cannot be punished but all the benefits availed because of the plagiarised thesis or research paper would be withdrawn, [an anonymous HEC official] added. The officer pointed out that the research paper was written in 2003 when there was no policy about plagiarism. “Dr [L.] just supervised that research paper so action can only be taken against [M.N.],” he said.
 L. is quoted in a 2014 article in the Express Tribune:  
[L.] added that he did not contribute to the piece and only provided the data the [N.] needed. [L.] also said that the case did not fall under the HEC plagiarism policy as it was published in a magazine, adding "I never benefited from it and the co-author included my name without my knowledge."
According to the Tribune, the article was on his CV in 2014, but there is no trace of the paper in his online CV today. In an article he published in 2014 ("I am not a plagiarist") L. writes
In 2003, I helped a faculty member, [M.N.], in collecting data for a report that he was writing on electronic voting. He subsequently published the report, in 2004, without my knowledge, in a non-recognised journal, of which he was the editor and used my name as co-author. Since it was typical that authors write names of advisers in their papers, and since I was his adviser in his other work, he put my name on the paper as an acknowledgment. No one, including me, could have verified the originality of anyone’s work then, since no anti-plagiarism softwares were available at that time. The paper eventually got listed in my CV — an innocent oversight as there was no vested interest involved because the journal was not of a notable stature even by Pakistani standards.
He goes on to state that what the Express Tribune has written is misleading and quotes another academic as calling this a conspiracy, because L. has been vociferous in denouncing Pakistani fake degrees and fake universities (which he, by the way, has). And says that it isn't a problem anyway, as there was no Pakistani policy on plagiarism before 2007.

I'm confused.

L. has apologized for plagiarism in a paper he didn't write that is no longer available at the journal named but where his co-author was editor and it still ended up listed on his CV by mistake? And now it has been removed from the journal page (and is not mentioned in the CV) but it is still on arXiv? Because it was put there before the Pakistani policy on plagiarism came into effect?

This has nothing to do with there being no software around to find plagiarism (not "verify the originality") -- you only put things in your CV that you wrote, and if you wrote an article, you know that you didn't copy it from somewhere else, so there is no need for any sort of software.

It doesn't matter if the journal is "non-recognized": Software filtering systems will find them and they will be available to others. If a PDF can be obtained online for a paper that has been retracted and there is no retraction watermark on the page, other scientists will be misled into thinking that the paper has not been withdrawn. People don't compare papers they are quoting with the CVs of the authors, as far as I know. And just deleting articles that have been found to be the results of academic misconduct may remove the offending article from the reach of software, but if anyone has quoted this article, there should be a proxy page that informs the reader what's up. 

Thanks to @gwarynski_ for the Express Tribune link, although it ended up being far too much research necessary to try and understand what happened! 

Sunday, March 27, 2016

New Approaches to Academic Misconduct in Denmark and Sweden?

There have been a number of high-profile cases of academic misconduct in the past few years, both in Denmark and in Sweden. The Swedish government has just issued a directive requesting that an independent examiner look at the necessity of changing the rules for investigating cases of academic misconduct in research. They request that a proposition be made for a timely and legally secure process for dealing with accusations of academic misconduct. ("En särskild utredare ska undersöka behovet av en ny hantering av ärenden som rör utredning av oredlighet i forskning och lämna förslag som säkerställer en tydlig och rättssäker hantering av misstänkt oredlighet.").

This comes on the heels of news (Retraction Watch reports) about a Swedish researcher who has been dismissed from the Karolinska Institut on multiple charges of academic misconduct

Denmark is a bit further along in the same process. They have had quite a number of scandals, so the UVVU (the Danish organization that looks at accusations of academic misconduct) has already prepared their own suggestions. They have a page with a number of links, and a relative thorough collection of the current practices in seventeen countries:Australia, Austria, Belgium, Canada, Denmark, Croatia, Finland, Germany, Great Britain, Ireland, Luxembourg, the Netherlands, Norway, Poland, Switzerland,  Sweden,  USA. Unfortunately, the report appears to only be in Danish. 

ScienceNordic has a nice overview of the the major scandals in Denmark and Sweden, including many links, in English.

The problems arise when lawyers are called into scientific disputes and judges decide what is and what is not good science. I think we need a sort of "Godwin's law" in science. If you involve a lawyer, you lose the argument. We need to focus more on peers discussing the science, although there do need to be sanctions for those found to have committed academic misconduct.

The Danish report lists the wide spectrum of possible sanctions found in the various countries (p. 20-21):
  • Issuing a correction
  • Reprimand
  • Supervision of future research
  • Suspension from scientific work
  • Retractions
  • Disciplinary sanctions such as being put on probation for future academic work
  • Rescinding of academic titles
  • Rescinding of the right to advise PhD students 
  • Withdrawal of internal resources
  • Repayment of research funding
  • No permission to apply for research funding, usually for a set number of years. 
The report makes it clear that there does need to be a system for appealing such a judgement and in particular the whistleblowers need to be protected.  It will be interesting to see what the Danish government decides to do and whether the Swedish report will be much different from the Danish one.

Update: fixed Goodwin -> Godwin

Thursday, March 10, 2016

German Defense Minister to keep doctoral degree

The Medical University in Hanover held a press conference on March 9, 2016 that was broadcast live on German television. They announced that the dissertation submitted in 1990 by the current German defense minister Ursula von der Leyen does indeed contain plagiarism (as documented by VroniPlag Wiki), but that they do not see an intention to deceive and thus they are not rescinding the doctorate. They see most of the plagiarism in the introduction, not in the results portion of the thesis. The president of the university stated that there are "errors, but not academic misconduct." The VroniPlag Wiki documents plagiarism on 27 out of 62 pages (43.5 % of the pages in the thesis affected), not only in the introduction, but also in section 3 (Thematic Background and Pathophysiological Fundamentals) as well as in the discussion.

RP online quotes law professor Gerhard Dannemann, one of the VroniPlag Wiki activists, as stating that this decision is irritating because plagiarism is academic misconduct, as has been decided time and again in the German courts when persons who had their doctorates revoked for plagiarism took their universities to court.

The Berlin daily newspaper Tagespiegel notes the close connections between von der Leyen and the MHH. Her husband is an adjunct professor at the MHH and director of the Hannover Clinical Trial Center GmbH that is affiliated with the medical school. She herself is a founder of the school's alumni association.

Since this case was published, VroniPlag Wiki has documented extensive text overlap in five additional dissertations (Acb, Bca, Lcg, Wfe in medicine, Cak in dentistry) and a habilitation (Mjm) from the MH Hanover. It will be interesting to see how these cases that affect people who are not politicians play out. In particular one would hope that these cases would also be dealt with in a timely manner and the results announced to the academic world.

In my opinion the MH Hanover has chosen what they think is a pragmatic solution. They split an academic publication into two parts, an important and a non-important part. Many biomedical researchers fall into the same trap: If the data is falsified or fabricated, they are quick to find fault, but do not find plagiarism to be a problem. This is, however, in direct contradiction to German court decisions that only see a dissertation as submitted as a whole. There is no "scientific core" that is okay, although the rest is tainted. The NSF in the USA is very clear on this topic:
„(1) fabrication, falsification, plagiarism or other serious deviation from accepted practices in proposing, carrying out, or
(2) Reporting results from activities funded by NSF; or retaliation of any kind against a person who reported or provided information about suspected or alleged misconduct and who has not acted in bad faith.“
This is how I see it: When entire pages are taken word-for-word without appropriate citation (for example the VroniPlag Wiki case Go with more than half of the pages containing plagiarism, among them 11 pages taken from the Wikipedia without reference) and passed off as one's own work, it is plagiarism and thus academic misconduct. It is also plagiarism (and thus academic misconduct) when throughout a text words or ideas are presented as the author's own when they are actually taken from another person. There is not a question of intent to deceive inherent in a definition of plagiarism, that can only have an effect on potential sanctions.

The MH Hanover deliberated and tried to find a way to have it both ways: The thesis contains plagiarism, but it is not serious enough to warrant rescinding a doctorate. I suspect this will provoke much discussion with current and future students who do not understand why they are given a failing grade for much less plagiarism.

Tuesday, March 1, 2016

Fake Academic Degrees in Russia

This is a guest post by one of the members of the Russian plagiarism documentation group Dissernet

Fake Academic Degrees in Russia 

By Andrei Rostovtsev  

The practice of awarding fake academic degrees to politicians, businessmen, doctors in clinics, professors in universities, and teachers in schools, that is, to all those who wish to use their new academic titles to step onto a faster career route, is widely accepted in Russia. The academic titles are awarded throughout the country. This business is based on the production of falsified dissertations. In early 2013 a group of five scientists and journalists established a social network called “Dissernet”. The Dissernet is a volunteer-effort free association aimed at making fraud and trickery in the awarding of academic titles transparent and well-known to the public. By 2016, Dissernet activists have identified more than 5000 plagiarized and falsified dissertations. In falsified dissertations not only is the text copied, but also the numerical data in it are assigned to a different year or region (in economics, law, and sociology), or to a different disease and treatment (in medicine), see discussion below. Over 1000 cases of such dissertations are documented on the website of the Dissernet (www.dissernet.org). Statistical data collected by the Dissernet yield a number of conclusions discussed below.

First of all, there is an important difference between the ways scientific writings are plagiarized in Russia and in the Western counties. In the West, the plagiarism is often associated with an intentional incorporation of other people’s texts or ideas in one’s own scientific research. That is probably why the ‘western style’ can involve many intricate small-scale mosaic plagiarisms intentionally placed in the original text. Yet in Russia, most often Dissernet deals with authors who have never done research and might have never even seen their dissertation texts at all. Such ‘dissertations’ are usually nothing else but a mere compilation of other people’s texts glued one paragraph after another in a haphazard way, something Weber-Wulff calls “shake & paste” [1].

In extreme cases the new text is just an older dissertation with a title page changed to reflect the new candidate. Sometimes the new candidate changes the subject of his or her ‘research’ too—usually by contextually substituting some terms throughout the whole text. For example, one notorious ‘scholar’ transformed a dissertation about the confectionary industry into a dissertation about the beef-and-dairy industry by substituting ‘dark chocolate’ with ‘homegrown beef,’ ‘white chocolate’ with ‘imported beef,’ and ‘nut chocolate’ with ‘bone-in beef ’ (see http://www.dissernet.org/expertise/igoshin.htm and http://cook.livejournal.com/202638.html, in Russian). In the meantime, all the data, tables, pictures, and spelling remained unchanged. Sometimes such authors also ‘update’ the dating of the statistics they refer to, thus making their ‘research’ seem to have been done more recently.

Detection of thousands of fraudulent dissertations by the Dissernet is mainly the result of a unique technology used. In Russia, along with the dissertation a so-called avtoreferat must be made publically available before the Ph.D. defense. The avtoreferat consists of a shortened dissertation content (usually 20–30 pages) and the main research results. Importantly, the texts of the avtoreferats are indexed by public search engines (such as Google or Yandex). The dissertation itself is not usually indexed, however. But if the dissertation contains large fragments of plagiarized text, as described above, its avtoreferat would also have text coinciding with earlier works. The specific Dissernet software is able to pick up the avtoreferats one by one and takes advantage of the search engines indices to look for textual coincidences within the whole publicly available corpus of Russian digitized texts, including texts of other avtoreferats. This program runs 24 hours a day and 7 days a week. Thus a few hundred thousand dissertations have been automatically checked. Furthermore, Dissernet takes advantage of the common practice of a chain-like fraudulent dissertation production. As soon as a rampant plagiarism is detected in one dissertation, it is very likely to be detected as well in other dissertations defended by the same dissertation council or with the same supervisor. This happens because the producers of fake dissertations in Russia work in a conveyor-belt mode using very limited sets of scientific texts as sources. By focusing on practically totally plagiarized texts, the Dissernet deals only with a small tip of scientific fraud in Russia. But even so, in problematic fields such as economics and law, about 3 % of dissertations contain large-scale plagiarism. In pedagogy this fraction is a bit higher, but still below 6 %.

Such large-scale dissertation fraud in Russia is a result of corruption that has paralyzed the whole system of awarding academic degrees: from dissertation councils established by the leading universities, where the PhDs are awarded, through the Higher Attestation Commission—the agency which coordinates and validates the awarding of academic degrees—and finally, what is also very important, to editorial boards of scientific journals, where scientific papers of the prospective doctoral candidates have to be published prior the defense. It is obvious that if no real research is done, then no relevant scientific papers could be published by such research. Clear affiliations of Russian scientific journals with the fake dissertation industry run by universities (more exactly, certain dissertation councils) have also been traced by the Dissernet. Those three cornerstones (dissertation councils, the Higher Attestation Commission, and journal editorial boards) are the necessary working parts of the mechanism running the conveyor belts of the academic fraud in Russia. Very often the same persons serve in these three cornerstone bodies at the same time.
Figure 1. Statistics on false dissertations broken down by scientific fields. (Dissernet data)
Figure 1 shows statistics of fraudulent candidate (Ph.D.) degrees awarded in different scientific fields based on the present Dissernet data (n=5215). As one can see, the most problematic areas are economics, pedagogy, and law. These same areas are the most problematic ones in the everyday (non-academic) life of Russians as well. In my view, this correlation is not accidental. The academic community naturally erects a barrier in the way of fake sciences and mythifications, which could otherwise define a climate for the life of whole society. In the areas, where the academic community is strong enough to resist the fraudulent practice of awarding fake academic degrees, the entire non-academic society is not driven by the false ideas. In addition, according to SCOPUS, the proportion of fake dissertations in each scientific field is inversely proportional to Russia’s international input in these disciplines [2].

Figure 2. Geographic location of the major universities producing fake dissertations.
Relative contributions into the total productivity for Moscow
and St.-Petersburg are given in percentages.
Figure 2 presents the geographic locations of universities that award the fake degrees according the present Dissernet statistics. Obviously, Moscow and Saint Petersburg play the most important role as they are among the largest cities. Other cities and towns fall behind. The scale of falsifications in the Caucasus region is relatively large but on the whole, their share in national statistics isn’t that high. This means the phenomenon of scientific fraud in Russia is not a marginal one. It is not localized somewhere on the outskirts of the country. Today it plays a role of an institution that is well integrated into the contemporary Russian state. Why do we qualify this phenomenon as institutional rather than a subject to free market?

Several recent laws and decrees protect owners of falsified academic degrees. The most important one (see http://www.rg.ru/2013/10/01/stepen-site-dok.html and http://www.saveras.ru/archives/6450) makes it impossible to strip a person of an academic degree if its defense took place before 2011. The authorities are quite reluctant to revoke the fake academic degrees, even if the defense has happened after 2011. The reactions from those accused of plagiarism by Dissernet varies from ignoring it, through calling it nonsense and accusations that it is politically motivated, to accusing Dissernet members of unprofessionalism and arguing that only appropriate dissertation councils have the right to assess the quality of dissertations (E. Denisova-Schmidt, personal communication). This point of view is broadly supported by state-owned mass media. Still, as of today, the Dissernet has managed to convince dissertation councils to revoke about one hundred fake academic degrees.

Last but not least, Dissernet investigations are relevant not only for an assessment of the Russian fraudulent academic world. Most importantly, the Dissernet provides a unique view on the deterioration of some institution’s reputations in Russia. In order to illustrate this point, several reference groups may be considered: members of the Russian Academy of Science (RAS), directors of Moscow’s primary and secondary schools, chancellors of Russian universities, regional governors, and members of the State Duma. Members of each group are selected if they have been awarded an academic degree during the last 15 years. Dissernet did not detect any falsified dissertation by the RAS members. Of 141 dissertations defended by directors of Moscow’s primary and secondary schools, 23 satisfied the Dissernet criteria for largely plagiarized texts. This amounts to 16 %—a rate which is more than three times higher than the probability of finding large-scale plagiarism in a random pedagogical dissertation.

Figure 3. Breakdown of fake dissertations by occupation: a reputation crisis.
This implies a silent mechanism at work selecting and supporting those who are prone to falsifications. The next group is chancellors of Russian universities, which has shown an even higher fraction of 21 %. Of that, one third of such universities are in Moscow. The proportion of politicians representing regional governors and members of the State Duma is even higher, reaching 41 % for the latter. In short, Dissernet performs a sort of a litmus test, identifying those dissertations prone to fraud and trickery, depending on the circumstances, and demonstrates the reputation crisis in Russia. This is illustrated in Figure 3. Why are the authorities, which are charged with larger responsibilities, subject to this stronger negative selection? This question will have to be answered by sociologists rather than Dissernet.

Despite aggressive state politics directed at the Dissernet, this public initiative has gained a good reputation and respect in Russian society in general, as evidenced by several awards and the fact that the name itself has become a meme.

[1] Weber-Wulff, D. (2014). False Feathers: A Perspective on Academic Plagiarism. Heidelberg, Berlin: Springer. 
[2] Rostovtsev, A. (2015). Some Observations on the Subject of Dissertation Fraud in Russia. HERB: Higher Education in Russia and Beyond, 3(5), 17–18. Available at https://herb.hse.ru/data/2015/09/22/1075563638/HERB_05_view.pdf