An application for reducing cheating cases in exams: Sample of Bartın University

In higher education, students engage in a structured and deliberately planned learning process to acquire both knowledge and skills. When it comes to evaluating their success, traditional assessment methods, predominantly based on exams, are widely used. However, the presence of academic dishonesty, particularly in the form of cheating, significantly undermines the integrity of this fair evaluation process. The aim of this thesis is to create an exam seating arrangement that minimizes potential cheating incidents by reducing physical proximity among students who have friendship ties. Thus, firstly, data were collected through a survey to determine the friendship ties among undergraduate students actively enrolled in the Department of International Trade and Logistics at the Faculty of Economics and Administrative Sciences during the 2023-2024 academic year. Using this data, a weighted student relationship network was created for the first time to represent the level of friendship among students. Secondly, seating arrangement data for a sample classroom where the surveyed students would take their exams were collected. For this classroom, a potential cheating scenario was assumed based on the likelihood of students sharing information during the exam with those seated directly beside them (to their right or left) or in front of or behind them. Finally, using these two data, an optimization model was developed within the scope of this thesis to address the exam seating arrangement problem for the first time. This thesis is, therefore, the first study to apply Operations Research (OR) techniques to the exam seating arrangement problem. The significance of the optimization model lies in its ability to guarantee the best seating arrangement, in contrast to the methods found in the literature. In the performed comparison testing, the algorithm proposed in the literature failed to produce the optimal seating arrangement even for a small sample, whereas the optimization model presented in this thesis produced the best seating arrangement for the same sample in a short time. Furthermore, optimal seating arrangements in which no students could cheat were achieved in a short time for the two classrooms examined within the scope of this thesis. Thus, the implementation of the exam seating arrangement developed in this thesis helps both to create a fairer academic environment for students and to reduce the physical and psychological burden on proctors.