Predicting strangers’ cooperativeness in Face-to-Face vs. Video-to-Video interactions: A case of inaccurate social prediction in mediated communication

In modern organizations, video conferencing platforms are frequently used for key interactions, such as job interviews and virtual team collaborations. However, the accuracy of social predictions during brief virtual encounters remains uncertain. This study investigates the impact of different communication media on the accuracy of predicting cooperative or generous behavior among strangers after brief interactions. Grounded in Media Naturalness Theory (MNT) and evolutionary psychology, we explore whether video-to-video (VtV) interactions affect the accuracy of predicting cooperativeness compared to face-to-face (FtF) interactions. Across two behavioral studies, participants engaged in FtF and VtV interactions, with the second study introducing a condition where eye contact was enabled using gaze correction technology (VtV^g). Our findings consistently demonstrate that FtF interactions lead to significantly higher prediction accuracy than VtV, where accuracy levels did not surpass chance. Interestingly, enabling eye contact in VtV^g did not significantly improve predictive accuracy. These results underscore the limitations of current video communication technologies in replicating the social cognitive and perceptual capabilities present in FtF interactions, with implications for remote interactions in organizational settings.