Monday, April 25, 2016

Seven more retractions for Danish computer scientist

Back in 2012, the German plagiarism documentation platform VroniPlag Wiki published a documentation about extensive plagiarism in a computer science dissertation submitted in 2007 to the Danish University of Aalborg at Esbjerg. This sparked some media attention (and was reported in this blog in May 2012) and eventually an investigation of the Danish national academic integrity body UVVU (Danish Committees on Scientific Dishonesty) was initiated. That body ruled on December 2, 2014: (English translation)
The Committee finds that the Defendant has acted in a scientifically dishonest
manner in the form of plagiarism [...]
I managed to obtain many of the documents produced by means of a freedom of information act inquiry. The UVVU mounted an exhaustive inquiry, and also informed the university currently employing the author as a professor of their decision. Interestingly, he is still listed at the university as of today, and has current publications listed.

The thesis borrowed heavily from journal articles and conference papers published either alone or in collaboration with others that turned out to include much text overlap with publications of other researchers. And after the thesis was defended, many more papers were published, again with others, that again contained extensive text overlap both with papers by other authors and with text from the dissertation. The true sources were about identifying criminal networks, the copied papers were on the topic of identifying terrorist networks.
The source is on the right, the edited copy on the left
VroniPlag Wiki lists more than 20 papers to date that are affected by substantial text similarities. The publisher at Springer and IEEE were informed, and this blog discussed some of the papers in June 2012.  In January 2013 eight papers were retracted by the IEEE.

Springer published 10 of these papers, but was quite indecisive as to how to deal with the situation. In a journal, a retraction can be published in the next available issue. However for conference proceedings, there is often no "next" volume in which to note the retraction. Of course, since the papers are all online, they can at least be retracted there. In January 2014 I found that Springer had published a retraction of one of the papers, but then retracted the retraction just a few days later, publishing an erratum instead.

During an idle search in April 2016, one of the VroniPlag Wiki researchers was surprised to see that Springer had quietly retracted seven of the ten papers. Of course, Springer wants the general public to invest $ 30 to read the retractions:

I was able to obtain four of the seven retractions because they were published in proceedings that my library has access to. The notices read as follows:
The publisher regrets to announce that the following chapter entitled [...] has been retracted. This chapter contains a large amount of reused and uncited material that was not published within quotation marks.
Looks to me as if Springer has come up with a new euphemism for that nasty P-word.

I find it troubling that Springer needed so many years to act on the information given to them about the problematic publications. And even though Springer is a member of the Committee on Publication Ethics (COPE),  they did not follow the advice given in the COPE flowcharts for dealing with such situations. This includes as a final step "inform the person who originally raised the concern."