ABSTRACT
The outbreak of the COVID-19 pandemic in 2020 required higher education systems to quickly adapt to a distance learning environment, which was facilitated by the deployment of the hybrid (traditional and innovative) learning model. However, the development of online written exams has been the most difficult challenge due to the commitment to ensure that the quality objectives of student assessment are reliably met. This work analyses the evolution, and the relationship of grades in written exams to determine the characteristics of a suitable and reliable remote exam. The data used were collected from the subject of Decision Methods for management science. Additionally, a global effect of the pandemic on the students performance was analyzed. For this purpose, Bayesian machine learning methods of classification were employed. Empirical results show that a proper design of the online exam should combine the following elements: adjusted time, conceptual questions, total randomization, biometric monitoring systems and additional control elements. The results also indicate that, in general, there has not been a negative global effect on student performance, concluding that the most important variable for determining the final exam grade is the type of question. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.