Moodlewatcher: one year experience of detecting and preventing fraud when using Moodle quizzes

Abstract

The possibility of obtaining results that have been fraudulently interfered with always exists, in all processes of student testing and evaluation. In traditional methods of evaluation, such as exams carried out using pen and paper, any fraud that occurs is extremely difficult to detect after it has taken place. The use of e-learning environments such as Moodle, however, may bring benefits that can be applied to the evaluation process in response to this problem. When using Moodle, all actions carried out by participants (regardless of whether they are teachers or students) can be automatically logged by the system, and when carrying out online quizzes this detailed information can be used as an aid in the detection of non-authorized situations that occurred during the process. The aim of this article is to present a tool, and the results of its usage over one year period, developed by the FEUP e-learning team, that permits the auditing and visualization of these situations in quizzes carried out on the Moodle@FEUP platform, thereby permitting the identification of potential offenders in this area. We also present an approach that helps to mitigate some kind of Denial-of-Service attacks to Moodle infrastructures based on Nginx, Apache and MySQL. Taking on board the most recent strategies and contexts of testing and evaluation, structural alterations to Moodle are also proposed which will enable the implementation of mechanisms to prevent these new methods of fraud.