What proportion of your assessment items are being text-matched?

The extent of non-original language that’s acceptable within a work is highly variable. In the instance of Law, where you want students to extensively cite legislation or decisions, anything less than 40% might indicate the student is giving you an opinion piece rather than genuine analysis of the law.

Conversely, the alarm bells go off for me when the rating is 0%. That can be indicative of a paper from a paper mill.

When setting up a Turnitin assignment you can nominate whether or not the reference list is included, I always include it, so that guarantees about a 10% non-original language match, depending how large the work is and what sort of analysis we have asked the students to make. In the Turnitin setup you can also specify whether or not direct quotes are included in the non-original language tally. Again, a very high percentage rating here can be indicative of students being over-reliant on direct quotes. This then is an instance of poor technique, not academic dishonesty.

On our Moodle sites we give our students a three or four page outline of why we use Turnitin and how it is used. In common with maybe 80% of UNSW academics, I both allow students to view their originality reports and to make multiple submissions of their assignments so that they can pick up instances of poor referencing and correct those poor references. In Turnitin there is a 24-hour delay on the generation of secondary and subsequent reports. This helps stop students “gaming” the system, tweaking the language until they think it is down to an acceptable level.

Finally, to answer the question about what proportion of your (institution’s) assessment items are being text-matched? my guess is that the common, acceptable incidence of non-original language probably is set somewhere in the range 15% to 30%. I am thinking of business and humanities studies when I state this. Engineering and IT studies no doubt differ, and probably have limited scope to use Turnitin anyway. Also, this is predicated on the reference list being included in the non-original language matching tally, and direct quotes also being included.

Adapted from a posting by David Meacheam – 19 November 2014 – in response to what proportion of your (institution’s) assessment items are being text-matched? David is studying for an MPhil in Higher Education, focusing on academic and student views of Turnitin/originality checking software use and is an academic at UNSW Canberra.

Leave a Reply

Your email address will not be published. Required fields are marked *