Measuring visual attention with 360 degree video stimuli in virtually immersive environments: a case study of police officers’ decisions to shoot

Abstract When 360-degree videos are viewed through a virtual reality (VR) headset, participants are exposed to remarkably realistic analogues of real-life scenarios with respect to both vision and vection. Embedded eye-tracking technology within our VR headset enabled the collection of visual attention data as participants viewed videos of volatile suspect-encounter scenarios recorded from the perspective of a police officer, and made lethal force decisions. However, there remains a gap in knowledge regarding how a researcher should map these attentional data onto areas of interest (AOI) within their 360 videos. Unlike 2-dimensional fields, AOIs within 360 environments cannot be defined using x and y coordinates. Thus, our solution was to develop a “roll-out method” in which 360-degree video stills are flattened into equirectangular shapes before defining AOIs. We then modified and developed algorithms for determining participants’ attentional hitting points (the location of the eye gaze in the 360 environment) using spherical coordinates, which are then used to determine attention to AOIs. We provide step-by-step and video instructions for our “point and click” application of this technique, directed towards an audience of social scientists with little or no previous experience with coding or computer science. We conclude with a case study application of this method to our own data, and found visual fixation on armed suspects mediated the relationship between high levels of physiological stress and slower reaction times in lethal force decisions. The discussion focuses on future application of our roll-out method for more general application to the social sciences.