On 22/04/2010 14:43, King, Martin wrote:
<snip>
> I have suggested to Turnitin through the channels provided by
> Northumbria Learning that, in the Inbox view, the Similarity Index could
> be augmented with the number of resources with which a student
> submission has similarity. For example 20% similarity from (20
> resources) or 18% (2 resources) - two similarly ranked reports with very
> different attributes you will agree. This would make it quicker to
> order the inbox and scan the headline data without having to look at
> every report.
>
A recent Academic Conduct Officers Forum at Oxford Brookes identified
what appears to would be a significant problem with this approach even
if TII implemented it. It appears that when TII searches for matches,
(a) it reports the first match it finds and (b) it searches its own
database before venturing out into the web, etc. With the increasing
number of pieces of student work in TII's database, more and more of the
matches nowadays are to other student work. When lots of students copy
from the same single piece of source material, perhaps taking different
slightly bits of it or paraphrasing bits slightly, what was in fact a
wholesale cut and paste from the original source often appears as a
large number of short snippets each bit taken from a different student's
work. An inexperienced eye might then decide it was very unlikely that
the submitting student copied from all those different students from all
those different universities and pass the work, completely missing the
underlying actual copying.
While this can be overcome by excluding other student work from the
data-set when seeking matches, such an approach prevents TII identifying
collusion with/ copying from other students' work within the cohort or
elsewhere... Is there a setting that can force TII to search out the
largest single match, or keep searching wider when a match is less than,
say, 20 words, or some other way of tackling this problem? Or is it
necessary to scan each piece of work twice, once against other students'
work and once with that work excluded (and, if so, how does one do that
when the students submit directly to/ through TII)?
All the best,
Jon Appleton
*************************************************************************
You are subscribed to the JISC Plagiarism mailing list. To Unsubscribe, change
your subscription options, or access list archives, visit
http://www.jiscmail.ac.uk/lists/PLAGIARISM.html
*************************************************************************
|