Applying Text Analytics to Examination of End Users’ Mental Models of Cybersecurity
Andreea Cotoranu, Li‐Chiou Chen · Journal of the Association for Information Systems · 2020
Mental models can explain how end users perceive their interactions with information systems, and inform cybersecurity awareness training. In this study, we used text analytic techniques to extract mental models representing cybersecurity concepts in learners at different levels of expertise. We applied these analytic techniques to text data collected from open-ended questions designed to capture learners’ understanding of cybersecurity concepts. We analyzed similarities and differences between learner groups using frequency, entropy and cosine similarity measures applied to n-gram features of their written responses. Our analysis showed that there is a difference in mental models between learners with informal exposure to cybersecurity topics and those with formal exposure. Furthermore, as a proof to demonstrate the predictive power of mental models, we correlated end users mental models with their perceived security. Finally, this study validated text analytics as a tool for capturing the mental models of end users without influencing the models.