You’re a Technion student who’s aced your exams and labs all year long. Are final exams really necessary — especially if the pandemic requires crowding into an exam hall on campus or holding them at home?
Professor Orit Hazzan of the Department of Education in Science and Technology and Ph.D. student Koby Mike just may have a viable alternative. Instead of testing the students, why not predict their exam scores using a machine learning algorithm?
The researchers would train a machine learning algorithm by feeding it data such as an individual student’s graded exercises, homework, and exam scores from previous semesters as well as achievement in other courses. The algorithm would then analyze the data to predict the individual’s final exam score without requiring them to take the exam. Students who prefer to take the final exam traditionally, in hopes of lifting their grade, could do so in socially-distanced safety as there would be significantly fewer students sitting for the exam.
The idea came about last spring when coronavirus cases started spiking and another lockdown went into effect. Final exams were to be conducted at home. To maintain the academic integrity of the exam, university officials used Zoom cameras and other means to supervise the test at home. But their methods were difficult to implement and often ineffective.
Hazzan and Mike acknowledge questions as to how reliable their model would be and whether it is ethical to predict a student’s grade without requiring a final exam. But at this stage of the research, their objective was limited simply to proposing a radical solution to the exam problem. With that in mind they said: “As the coronavirus has transformed teaching in a way that will most probably change the future of teaching, thought must be given to ways in which the evaluation of learning … may lead to a change in future evaluation methods.”