Beyond camera presence: individual differences and emotional climate shape conversational memory in online meetings

Online meetings have become ubiquitous, yet whether keeping the camera on improves or impairs memory for what is said remains contested. The cognitive-load (Zoom-fatigue) framework predicts impairment, whereas the social-presence framework predicts facilitation. We adjudicated these competing predictions in a within-subjects experiment in which 86 university students conversed dyadically with confederates on a custom WebRTC platform across four conditions: camera-off, and camera-on with positive, neutral, or negative emotional climates. After the conversation, a surprise free-recall test measured memory for the partner's spoken content. Linear mixed-effects modeling (marginal R ² = 0.15; conditional R ² = 0.34) revealed that recall was reliably predicted by sex (β = -0.21, p = 0.006) and agreeableness (β = -0.20, p = 0.006), but not by camera presence. Extraversion further interacted with the neutral climate (β = -0.13, p = 0.021), such that extraverts recalled less when their partner displayed no interpretable socioemotional cues. These findings argue for foregrounding individual differences and emotional context, rather than camera state alone, in models of memory in videoconferenced communication.