Change gently: an agent-based virtual interview training for college students with great shyness

Wenxiu Geng, Chao Zhou, Yulong Bian

Abstract: Virtual Reality (VR) training systems are gaining popularity for psychological and skills training. Shy college students may perform poorly in interviews due to a fear of being evaluated, anxiety about real interview scenarios, and a lack of interviewing skills. In response, a pre-training environment (VR training system) was designed and developed, combining Virtual Reality Exposure Therapy (VRET, simulating a safe and controlled real interview environment) with biofeedback therapy (intuitive multimodal assessment and feedback). Virtual agents can play a positive role in psychological intervention and skill training. Different personalities of interviewers may help shy individuals overcome interview anxiety and enhance their interviewing skills. Therefore, we created virtual interviewer agents with three distinct personalities based on Eysenck’s personality theory. We improved on existing methods for designing agent personality traits by defining expressive traits (e.g., different physical, behavioral, and verbal traits) and reactive traits (differential evaluative feedback based on the user’s mental state). This self-training system requires no additional labor and provides initial training for shy college students prior to actual interviews. Ultimately, we recruited and selected 16 shy college students who underwent three VR-simulated interviews using our system. Expert evaluation and physiological measurements were employed to assess their interview anxiety and performance. The results of the user study indicate that our agent-based VR training system effectively assists shy college students in overcoming interview anxiety and improving interview performance.