PROF DR AHMAD HASHEM is the General Manager (Chief Technology Offi cer) of METEOR Technology & Consultancy Sdn Bhd. One of his projects is to implement a system called “similarity checker” as part of OUM’s efforts to combat plagiarism. Let’s here what he has to say about it.
Harvinder: Plagiarism is considered a serious problem in almost all learning institutions worldwide. I understand OUM is not spared this problem too and has lately been using a system called “similarity checker”. Please tell us more about this system.
Prof Ahmad: A similarity checker is not a plagiarism detection system. All plagiarism detection systems compare each submitted essay against internal and external sources. The external sources include freely available material on the Internet, digital subscription databases of reference material and other proprietary sources. Plagiarism detection systems, such as Turnitin, work on a subscription plan where a fee is charged on each essay submission.
The similarity checker, on the other hand, compares essays, assignments and reports in a database against each other for similarity. It does not compare with external sources. The essays for a particular course can be compared with all essays in the current semester or with all essays in the database.
We use a system based on the free open source program called WCopyfi nd, developed by Lou Bloomfi eld who is a professor of Physics at the University of Virginia. The original program was designed for manual operations, whereas in OUM’s case, with our large student numbers, the system must be automated and integrated with a database. Thus, the original program has been modifi ed and extended to meet the current needs of OUM. More information on WCopyfi nd can be obtained at http://plagiarism.phys. virginia.edu/.
Harvinder: Why has OUM preferred this system over the others and what are the benefi ts?
Prof Ahmad: The major compelling reason is that this system is free and can be customised to meet the exact needs of OUM. The usual subscription-based plagiarism detection system can be expensive for a large operation such as OUM. For instance, Turnitin charges RM7.00 per essay submission to be checked for plagiarism. Multiply this amount by the number of essays and it is easy to understand that free is much cheaper.
The other main reason is that a similarity checker system is faster than a subscribed plagiarism detection system. In a pilot trial, we found that our system can compare 1,000 essays contained in a database in less than one minute. On the other hand, the subscribed detection system takes a long time because comparisons are made against varied sources such as the Internet, databases of books, journals, and so on. A typical OUM course can easily take more than 24 hours for a plagiarism report to be produced.
Therefore, given that we have to check hundreds of thousands of essays, the plagiarism detection system is neither practical nor cost-effective.
Harvinder: What types of assessment can be scanned for plagiarism using this system?
Prof Ahmad: Any form of text document such as assignment essays and reports can be submitted for checking.
Harvinder: How does this system function in detecting plagiarism and how reliable are the detection results?
Prof Ahmad: This system does not detect plagiarism as such but highlights the extent of similarity among essays that are compared. Higher similarity between two essays means that the essays share many similar passages. The result of the comparison is shown in a report that presents the similarity index, which is the extent of similarity in percentages of the total text between every pair of essay being compared. The assessment of similarity is very accurate.
Harvinder: In your opinion, will this similarity checker system discourage learners from plagiarising in their assignment? To what extent will the system prevent plagiarism?
Prof Ahmad: We want to encourage learners to produce original work and to cite all works that were used as reference. Proper academic standards must be adhered to at all times. Writing an essay should be a learning experience. Research, analysis and thinking must take place.
The aim of using this system is to identify essays that may not be original and separate them from the rest of the documents in the database. It may not prevent plagiarism absolutely - no system can - but it is hoped that with the system in place, it will help to advance honesty and scholarship accordingly.
Harvinder: Will the similarity checker system be viable for future use in OUM or will it be outdated in time?
Prof Ahmad: Because this system does not rely on external sources and all essays are compared against each other within the internal database, it is apparent that this system will be useful as long as the comparison can take place.
Harvinder: What are OUM’s future plans with regards to the use of the similarity checker system?
Prof Ahmad: This initiative on our part is still in its infancy. However, it is hoped that new approaches and strategies could be developed based on the learning experiences and lessons learnt from using this system.