Examining alternative approaches used for identifying test cheating in online exams

Test cheating is a very comprehensive concept that includes various sub-dimensions. In tests that include sub-factors such as having prior knowledge of the item, cheating, collusion and test falsification after the test, fraudulent behavior can be defined as any action taken to gain unfair advantage before, during or after the application of the test. In recent years, with the spread of large-scale online exams, the need for new methodological approaches to detect test cheating has increased. This study proposes new approaches as an alternative to psychometric-based methods such as person-fit statistics, copy-similarity measures, and response time models. In addition, it is thought that by using the strengths of supervised classification methods, it may be possible to reduce the rate of false evaluations in detecting cheating behaviours. It has been shown that the proposed approaches provide better cheating detection rates than other classical methods described in the literature. In this study, new approaches are proposed as an alternative to psychometric-based methods such as person-fit statistics, duplication and similarity measures, and reaction time models, in which these methods are considered together. In this context, it is predicted that by using the strengths of the supervised classification methods, it is possible to reduce the false evaluation rates in detecting cheating behaviors. The proposed approaches have been analyzed using a ready-made data set and it has been shown that these approaches provide better fraud detection rates than other classical methods described in the literature.