This is a UKEdChat article, supported by Urkund.
Towards the end of the academic year, it is time for many students to hand in their long-awaited assignments, signalling the end of another semester of inspiring teaching and learning. For their teachers, being faced with a collection of essays can seem overwhelming and being able to check the originality and authenticity of each one could take a considerable amount of time.
With the continuing evolution of online sources, the temptation for some students can be to copy and paste elements of text into assignments. Backed by a strong argument, recognising other sources is completely acceptable, but when no reference is made to the source, and the work is claimed to be original, that is where accusations of plagiarism come to light. Most students are honest and completely trustworthy, and when signs of plagiarism are detected, it may be a sign that they are struggling with aspects of their life or learning. Indeed, Jamie Heywood – a Course Manager at Bedford College (UK) – told us, “Plagiarism is a growing issue in education, with the rise of essay writing websites and the vast array of readily available online sources accessible to students who fall into the trap of copying and pasting.”
The Urkund Experience
There is a marketplace now of online plagiarism checkers, but the technology behind Urkund stands out, as it is innovative in what it reports along as being intuitive for users. Founded in 2000, the Swedish company has built a platform that goes through an analysis process which, in its simplest form, can be described by a series of steps:
- Text extraction and data mining
- Source retrieval
- Text alignment
- Report generation
During the first step of the analysis process, the text from an uploaded document is extracted and the system tries to learn from it by looking at different classifiers in the text. It’s then indexed in order to make it searchable for later use as a source for potential plagiarism. The purpose of source retrieval in the analysis process is to find as many potential sources of plagiarism as possible. For that, three major source areas are considered: published content, previously submitted documents and the Internet.
The source retrieval step is crucial from a computational and a quality perspective. The number of sources is reduced to a subset of documents so that text alignment can be as effective as possible.
The text alignment part of the analysis process is where the submitted document and potential sources go through a set of advanced algorithms to find potential plagiarism.
Finally, a report with findings is generated and sent to the user, allowing the teacher to determine if plagiarism has occurred or not. Crucially, the platform offers grading and feedback through integrations with school, college or university internal systems.
We asked a couple of college teachers to try out the system, armed with a couple of sample student’s work to allow the system to analyse.
Review 1 – Jamie Heywood – Course manager, Bedford College
My first thoughts when visiting the website was that it appeared visibly slick and easy to navigate. After logging in, I was taken to a slightly more humble interface with the ability to upload documents, create individual folders and a link to a quick reference sheet (which I would certainly recommend reading prior to use). I was able to upload my first document effortlessly, with a one-click file upload box (with the ability to add multiple files in bulk if needed) which processed instantaneously.
Under thirty seconds later, I obtained the originality report with a link to the full analysis. The similarity percentage was included in the email report for those instances you need a snapshot judgement.
When viewing the full analysis, there was a clear overview page (with the similarity percentage, submission details and a summary of findings) as well as the option to view the entire document. This broke down the submission further with the options to view the individual sources and findings as well the full document with a similar text highlighted. Switching tabs to view this felt unnecessary and could be amalgamated further, however, there were a plethora of useful features with the ability to not show quotes and brackets in the text, remove certain sources from the analysis and export the report as a PDF.
I decided to challenge the system by uploading a previously submitted plagiarised assignment (unfortunately, I had a selection to choose from!) to see if it would be able to detect the similarity. Urkund did this successfully with a full breakdown of the sources used and highlighting the large similarity percentage.
As an alternative to other popular plagiarism checkers, and for those wanting to try a new system with cool features, Urkund is certainly worth a shot.
Review 2 – Adam Chapman – Course Manager – Early Years and Higher Education
Urkund is presented professionally and allows the user to see where there are potential areas of concern in terms of plagiarism; I was keen to see the intricate mechanisms that would be beneficial for Teachers in the UK who are often constrained to heavy marking deadlines and heavy workloads generally.
Upon using the BETA version, I was pleasantly surprised at how useful the software is in identifying where the author of the submission had percentage scores for plagiarism. It is often an area of confusion for teachers who see high plagiarism scores, however, scores may be due to the fact that some databases have a high volume of the same paper, for the same topic.
In this case, Urkund highlighted where they had seen certain references in previous submissions. The BETA version of Urkund gives you options of seeing text that has high scores individually (Findings), looking at sources (references and application of) and the entire text. This is a nice addition and provides the user with easy to digest chunks of data that can be used when providing feedback to students. I particularly enjoyed the option of including in-text referencing to see where Urkund would pick up on any differences. The BETA version demonstrates a lot of promise. It highlights text as a step by step process, which is useful for time-constrained educators.
Urkund aims to present analysis available to educators in a different, more user-friendly format, and its ability to be integrated within internal IT systems is an additional bonus. Speed is of the essence and retrieving analysis within moments is also appreciated. The platform is working on expanding features and products to offer broader functionalities including the soon-to-be-released cross-language plagiarism detection, meaning potential plagiarism can be found even when translated from another language.