An experimental study of tie transparency and individual perception in social networks

Tie transparency, which may provide more unbiased information to deepen mutual understanding, thus builds trust and prompts cooperation in social networks. Little is known, however, about social connections' transparency. We introduce knowable degree (KD) to characterize the transparency of a social tie, defined as the number of other entities who perceive the tie. We design a two-phase experiment to collect KDs in a network of 155 students. We find that structural property and node attribute significantly predict tie transparency. Meanwhile, we also find there almost always exist a few covert ties due to non-reciprocity. Furthermore, we focus on exploring the boundary of scopes of perception and evaluating individuals' perceptual capability. We describe the two degrees of perception phenomenon that people can generally catch the relationships between their 2-neighbours at most. We propose a generic quantitative model to recognize high-capability perceivers, who are found more sociable and enjoy exploring the social context as well.