Flaws highlighted in a plagiarism-detection system used by the University have been explained by the Deputy Head of the Law School, Gordon Stewart.
In last week’s Salient, it was reported that British lecturer Simeon Yates used Turnitin to test his own work for copied material and found that it failed to recognise his submission was 100 percent copied from his own previously-published output.
Turnitin is an online system used by a number of faculties which checks submissions against both current and archived internet content, commercial databases of periodicals and previously submitted material.
Stewart believes that Yates’ research results were “merely sensational [rather] than valid.” “Unless Yates’ first piece [was] on the Turnitin database, this will happen,” he says, “[He] should have run his previously published piece through… and then later run his second piece through.”
Stewart added that he often ran work through the system twice and conceded that while Turnitin was not yet a perfect system, it would improve as its database grew.
Stephen Marshall, Acting Director at the University’s Teaching Development Centre, confirmed Stewart’s interpretation of Yates’ findings, and noted that Turnitin greatly reduced the work that University staff needed to do to document and resolve plagiarism problems.
[ssba]
The bichs imeges
>
Ah OK: if I understand correctly, we should instruct students to please first run the texts they plagiarize from through the system before they submit their plagiarisms?
I had almost exactly the same experience as professor Yates when I tested Ephorus, a European imitation of Turnitin. Everything went fine when I fed it the German text about Edison in http://www.forschungsstandort.ch/webexplorer.cfm?ID=1&tlid=1 : it recognised the forschungsstandort source, and the Wikipedia article forschungsstandort had itself plagiarized from.
But when I submitted the French translation of this text as in http://www.forschungsstandort.ch/webexplorer.cfm?id=1&tlid=2 , ephorus was unable to find any match on the web. A Google search for the sentence “Son mérite tient principalement au caractère commercialisable de ses inventions” gave http://www.forschungsstandort.ch/webexplorer.cfm?id=1&tlid=2 as the first and only hit.
Ephorus’ demo is free (non paying) until the end of October, so it was fun trying it. But what’s the point of paying huge fees for comparative plagiarism detection systems that perform worse than a search engine on Internet material, that cannot possibly access what gets exchanged in private forums, wikis, mailing-lists, e-mails and Google Docs pages?