Hide and Seek: Choices of Virtual Backgrounds in Video Chats and Their Effects on Perception

In two studies, we investigate how users choose virtual backgrounds and how these backgrounds influence viewers' impressions. In Study 1, we created a web prototype allowing users to apply different virtual backgrounds to their camera views and asked users to select backgrounds that they believed would change viewers' perceptions of their personality traits. In Study 2, we then applied virtual backgrounds picked by participants in Study 1 to a subset of videos drawn from the First Impression Dataset. We then ran a series of three online experiments on Amazon Mechanical Turk (MTurk) to compare participants' personality trait ratings for subjects (1) with the selected virtual backgrounds, (2) with the original video backgrounds, and (3) with a gray screen as a background. The selected virtual backgrounds did not change the personality trait ratings in the intended direction. Instead, virtual background use of any kind results in a consistent "muting effect" that mitigates very high or low ratings (i.e., compressing ratings to the mean level) compared to the ratings of the video with the original background.