Online vs. written multiple-choice questions tests: accuracy and usefulness

With the online resources becoming common, one can ask what the limitations are when applied to online examinations. Standardised multiple-choice questions (MCQs) test are commonly used either as a tool to enable student self-examination or as a tool to test large number of students in a more efficient way. However, the means of delivery and what one intends the tests to measure should be carefully considered. During last two years in teaching, the course 'Signals and Systems', both online and written, proctored and non-proctored, MCQs test were used in production environment (in vivo). Obtained data suggest that test delivery method strongly influences test accuracy when measuring the knowledge students gained. Online non proctored delivery, although being the simplest, yields strongly biased results and is suitable only as a tool to enable student self-assessment of the gained knowledge, while the online or written (non-strictly) proctored tests are still the method of choice if one wants to objectively measure the knowledge gained by students.