Analysis of Writing Data for Cheating Detection in e-Testing

Abstract

e-Testing is effective in relieving time and space limitations for examinations. However, a drawback is that user authentication employs only login credentials , making cheating easy. We examine variations in personal writing data in e-testing with the aim of detecting cheating. We assume two behaviors: answering questions by oneself and replicating others’ answers. We used the sub-stroke method for data analysis, which allowed confirmation of differences in the change over time in writing speed and pen pressure, and differences in average pen angle and direction.