Impact of Anchors’ Voice Characteristics on Viewer Engagement in Live Streams
(A2026-133288)

AUTHORS

Qingli Zeng, Hebrew University of Jerusalem; Sandeep Chandukala, Singapore Management University; Hai Che, University of California at Riverside; Lifeng Yang, Shanghaitech University

ABSTRACT

In the rapidly growing live-streaming market, anchors primarily communicate vocally while viewers engage through both immediate active participation (chats) and sustained commitment (subscriptions). Drawing upon dual-process frameworks of social approval and social influence, this study examines how anchors’ expressive energy, stressed vocal tones, and cognitive depth differentially affect these distinct engagement behaviors. Leveraging over 1,000 hours of gaming content from top streamers, we employ advanced audio, video, and text analytics to uncover nuanced engagement patterns. Results indicate that expressive energy and cognitive vocal expressions enhance sustained commitment without significantly influencing short-term chat participation, whereas stressed tones positively impact both engagement types. Furthermore, contextual moderators such as chatroom sentiment and channel popularity significantly alter these relationships, revealing that vocal cues affect viewer engagement differently depending on the social environment. This research contributes theoretically by delineating distinct psychological mechanisms underlying viewer engagement and offers actionable guidance for platforms and content creators.