5.5 Exam proctoring
Sometimes a high-stakes exam is essential, perhaps because it is a required component of professional recognition. This raises two significant issues. How can the institution be sure a completed exam paper is the unaided work of the student, and how can the institution be sure students did not have access to the information resources and support that would be available to them in everyday life?
In examination halls, these issues are addressed by the use of invigilators or ‘proctors’. They are responsible for checking student IDs, collecting mobile phones and other study aids, preventing communication and enforcing timekeeping. They are not infallible, but their actions make it highly probable that the exam scripts handed in are the unaided work of the named student.
In an online setting, some of the problems can be removed or reduced by good assessment design. Questions that test understanding or factual knowledge can be completed by students who have access to their normal technology and resources. Questions that require some sort of personalised response, for example quotes from contributions the student has made in the past, make it more likely that the individual sitting the exam is the same individual who worked through the course.
If an online equivalent of the examination hall is necessary, online proctoring provides a way of making this a rigorous process. There are two approaches: random proctoring and full proctoring.
- Random proctoring: Software is used to take pictures of the student at random times during the examination; these pictures are analysed automatically to check that the same person is pictured each time, and reports are sent to the examiner.
- Full proctoring: An invigilator proctors the exam using webcam footage. At the start of the exam, each student displays their surrounding environment, showing it is clear of study materials. The proctoring may take place live or by reviewing recordings.
In a 2020 report [Tip: hold Ctrl and click a link to open it in a new tab. (Hide tip)] that covers online assessment and verification practices, Habib and Sanzgiri outline a third approach, TESLA, which is not yet in regular use, but which combines tools already in use separately. This system includes:
- face recognition: as with random proctoring, still and video images are analysed and compared;
- voice recognition: recordings of the student’s voice made during the exam are compared with each other and with previous recordings;
- plagiarism checks: the text submitted is automatically compared with published material and with work submitted in the past;
- key-stroke patterns: patterns of press and release times for different computer keys are compared with previous patterns.
Although students are used to major restrictions on their behaviour and a high degree of surveillance in an examination hall, moving these practices to their home environment highlights how intrusive and problematic these restrictions can be. If you are interested in exploring some of the ethical problems connected with proctoring, including some of the problems in extending the use of systems designed for relatively wealthy white Americans, Shea Swauger’s (2020) post Our Bodies Encoded: Algorithmic Test Proctoring in Higher Education gives a comprehensive overview.