Wikipedia:Wikipedia Signpost/2015-02-25/Op-ed

Last October, I came across the Oxford Textbook of Zoonoses (2011) published by Oxford University Press (OUP). I noticed that chapter 31, "Marburg and Ebola viruses", contained a fair bit of text that was nearly identical, word for word, as that in the Wikipedia article Ebola virus disease. A page from the book may be seen on Google Books, with at least the "natural reservoirs" section being nearly verbatim and some parts of the rest of the chapter containing great similarities.

Initially, I made an assumption that someone had copied and pasted from this book into Wikipedia. However, thankfully we have the ability to go back and view every version of Wikipedia that has ever existed. I could thus determine that the content in question was added to Wikipedia back in 2006 and was subsequently edited and expanded between then and 2010, when the greatest similarities occur. From this I could conclude that it was partly written by the Wikipedians and.

Next, I wondered whether one of these individuals was the author of the OUP chapter, namely, Graham Lloyd of the Special Pathogens Reference Unit at Porton Down. I contacted the user who had made the majority of the contributions, who turned out to be a virologist in Australia who assured me that while he had contributed to Wikipedia, he had never contributed to the Oxford Textbook of Zoonoses.

Finally, I looked for attribution of Wikipedia in the Oxford Textbook of Zoonoses and a release of this book under an open license as required by Wikipedia, and the result was that neither of these have been performed. The hardcover version of the Oxford Textbook of Zoonoses retails for $375. I discussed this issue with the legal team at the Wikimedia Foundation, who contacted the Oxford University Press. We were hoping that they could negotiate both attribution and release under an open license.

The reputation of Wikipedia in academia often seems to be that it is good enough for academics to use and even occasionally claim as their own work, but not good enough for either students or the “unwashed masses”. Thus I believed that convincing one of the world’s foremost medical publishers to both attribute and use an open license would be difficult. The legal team at the WMF, however, was optimistic. Initial emails from OUP indicated that this case would take longer than usual, as the people involved were “all over the world doing important Ebola work”. This, of course, is not the first time we have come across the academic literature copy and pasting from Wikipedia. In 2012, I discovered a medical textbook had also extensively copied from Wikipedia. (Also see the Signpost's 2012 special report on the misappropriation of Wikimedia content.)

At Wikipedia, we are happy to work with publishers. A year or so ago, I helped guide the company Boundless, which creates open access textbooks mostly based on Wikipedia content for first year university students, on how to appropriately attribute. These books were already released under a CC BY SA license. We attempted to work with the OUP in the same fashion.

On January 20, 2015, the OUP acknowledged that the content originated from Wikipedia and agreed to attribute Wikipedia, but were having difficulty with the open licensing. Following further inspection of the Oxford Textbook of Zoonoses, I found more inconsistencies. For example, while parts of the text were exactly the same, the author had not consistently used the same references. The references used on the Wikipedia article supported the text, but the references in the Oxford Textbook of Zoonoses that were changed did not support the text in question. The question remains as to why the references were changed. As a result of these changes, the quality of the copied content was lowered.

On February 5, 2015, I emailed the OUP offering to rewrite and update the chapter in question in collaboration with fellow Wikipedians. The next day, they replied via e-mail stating that they had already “independently decided to update the chapter and that that work [was] already in hand”. Writing a textbook chapter takes a fair length of time, likely weeks rather than a few days. Looking at the time line, it is questionable whether the OUP ever seriously intended to attribute Wikipedia. While our content passed their review processes, they claimed it was simply an “inadvertent omission of citation”. It is likely that a replacement chapter was requested immediately after the WMF legal department contacted OUP’s team.

The one good thing that has come out of all of this is that Wikipedia’s content passing a major textbook publisher review processes is some external validation of Wikipedia’s quality.

A look at the references

 * Both Wikipedia and the Oxford Textbook of Zoonoses include "The absence of clinical signs in these bats is characteristic of a reservoir species. In a 2002–2003 survey of 1,030 animals which included 679 bats from Gabon and the Republic of the Congo, 13 fruit bats were found to contain Ebolavirus RNA". Wikipedia cites a 2005 article from Nature, which does support it.  The Oxford Textbook of Zoonoses cites a 2009 article from BMC Infectious Diseases, which does not support it.
 * Both include "no Ebolavirus was detected apart from some genetic material found in six rodents (Mus setulosus and Praomys) and one shrew (Sylvisorex ollula) collected from the Central African Republic". Wikipedia cites it to a 2005 article from Microbes and Infection which does support it, while the Oxford Textbook of Zoonoses cites [a 2004 article] from Emerging Infectious Diseases which does not support the content.
 * Both state "Of 24 plant species and 19 vertebrate species experimentally inoculated with Ebolavirus, only bats became infected" and both use the same reference, a 1996 article from Emerging Infectious Diseases.