A Markov Chain Approach for Modelling Normally Distributed Online Assessment Time in a University Setting

Masikini Lugoma, Masengo Ilunga, Violet Patricia Dudu, Amuli Bukanga

The Markov chain (MC) technique is applied to an online invigilated assessment situation to predict the writing time in a typical university setting. Different students' writing times cannot be determined accurately in advance and are associated with randomness. This preliminary study simulates data related to the time to download the question paper and the writing time, from a normal distribution. The time variable is simulated to have a reasonably good approximation of the real settings where most students’ writing times are spread around the expected value, namely the mean. Simulations are conducted based on the experience and knowledge of researchers in the online teaching and learning environment. Computer simulations demonstrate that the writing time estimates depicted a stable convergence, thus giving clear insights for optimising online assessment implementation. The findings showed that the average writing time of a selected trial reaches a stable value at 1.498 hours (89 minutes) within the confidence interval [0.6, 2.5], at 95%. Therefore, these results offered a more realistic range of feasible times to guide academic practitioners on the planning and implementation of invigilated online assessments.

Full Text