Ensure academic integrity in programming courses with Codio's plagiarism checker.
There are various forms of student code copying, ranging from desirable collaboration through to straightforward cheating. Codio is pleased to announce the launch of a best-in-class code plagiarism detection technology called Etector into its platform.
Why use plagiarism detection?
The reason to detect code copying is not to catch students and then name and shame them. Rather it is a highly useful way of understanding both desirable (collaboration) and undesirable (cheating) behaviour. Above average levels of cheating might suddenly happen because a module is not well structured or taught. Or one particular assignment is just a lot harder than you thought, and so students 'borrow' a little code from one another to get the assignment in on time. Understanding the integrity of students, and by extension the course and the institution, also matters and our plagiarism detection features are there to help with different aspects of quality, not to decide who needs punishment.
Etector detects twice as many cases of code cheating over standard tools.
Etector was originally developed for and privately released to a group of universities including Princeton University and the University of Pennsylvania.
Benedict Brown from the University of Pennsylvania says “We had been using other systems and were delighted that Etector is twice as effective at detecting cheating. We are delighted to hear that Codio has now integrated Etector and made it available within their platform.”
The plagiarism checker ensures lecturers and teachers of computer science courses make their own decisions on how to interpret results. However, the plagiarism detection report provides enough data and analysis for a lecturer to make a conclusive, final decision. The burden of project data preparation and submission to remote systems such as Moss and JPlag is removed, resulting in a single click process for the lecturer or teacher.
There are various plagiarism detection systems available such as Moss and JPlag. However, these systems were not developed for university programming courses. Therefore, they can require considerable effort to submit large files of student code projects and to interpret the results.
Vivek Pai, the creator of the Etector technology, was a Computer Science professor at Princeton University where he was unhappy with the results of existing plagiarism detection systems. "I found myself spending an awful lot of time, in short supply in CS departments, trying to interpret the statistical output of existing systems. Doing this for potentially hundreds of submissions was a real burden. I needed output that not only provided clearer rankings but also did a better job of detecting matches for any given pair or group of submissions. Both Princeton and other universities have been delighted at the improvements Etector has delivered."
Codio integration of Etector delivers even greater time savings thanks to its ability to manage the entire code preparation process.
“Etector is now seamlessly embedded into the Codio platform. This removes the need to manually prepare hundreds of student projects before submitting them to a remote plagiarism engine. Codio has effectively reduced the checking process to a single click.” - says Freddy May, Chief Product Officer
Codio is programming language agnostic and so supports almost all programming languages. There are special parsers for Java, C, C++ and Python that deliver a little more accuracy.
How it works
Codio has made the process extremely simple. Select a class and an assignment or code project. The image below shows students who have completed the Bubble Sort assignment. Select the ‘Plagiarism’ button and a new window will open with Plagiarism results.
On the next screen you can restrict file types to check. Select the 'Start' button to generate a new report.
The Plagiarism Detection Report
The final plagiarism detection report compares the student files. You can also check against other reference projects and code files.
The final report allows the lecturer to drill into projects that show similarities. Color coding and font-sizes are used to assist the lecturer in making decisions. The report highlights similar projects in a ranked lists along with various statistical pointers. The technology highlights suspected areas and their differences of a students code.
Alternatively, you can view suspect student projects side by side.