Analysis of YouTube Comments to Inform the Design of Virtual Reality Training Simulations to Target Emotional Arousal

Abstract

Workplace safety remains a concern in the construction industry as fatality rates continue to rise. While hazard recognition training programs have been implemented using multimedia-based modules, their effects have not been broadly reflected on construction sites. In an effort to provide realistic and engaging training, a recent focus on virtual reality (VR) for an immersive learning experience has been shown to offer benefits to improve traditional lecture-based training. Such virtual environments can be especially useful for simulating hazard recognition tasks that are inaccessible in real-life settings due to the potential dangers they pose for trainees. However, due to the focus on applications for performance assessment and procedural training, strategic elicitation of emotional arousal, which has been shown to be a precursor to desired learning outcomes in hazard recognition training, has not been explored for construction-specific VR applications. To guide the development of such VR environments that target emotional arousal for learning, this study used opinion mining to catalogue the features that yield or inhibit an emotional reaction in similar video simulations posted on a public video sharing platform (YouTube). Design insights such as the need to provide agency in the simulations, introducing nonplayer characters in the scene, and the like, are presented. Here the authors discuss specific implementation strategies derived from the study findings that developers can use to elicit emotional arousal in a construction-specific virtual environment. Virtual reality (VR) training programs are increasingly being used in the construction industry to improve hazard recognition and reduce fatalities. VR provides an immersive learning experience that can simulate tasks that would be dangerous for trainees to perform in real life. However, previous VR training programs did not focus on leveraging emotional arousal, which has been shown to be beneficial for learning outcomes. This study used natural language processing to analyze user comments and identify key features that lead to an emotional reaction in virtual simulations on YouTube, and provided insights and strategies for developers to create construction-specific VR environments that elicit emotional arousal.