ASNAC2017で発表を行います

2017年11月28日～29日にかけて開催されるThe 2nd Australian Social Network Analysis Conference (ASNAC 2017) (Sydney, Australia) で以下の発表を行います。

Oral Presentation

Igarashi, T, Kato, J, Shiraki, Y, Hirashima, T, & Tamai, R (2017). Chasing stars and confirming alliance: Two effective strategies for learning social network structure.

This study reports a novel technique that unveils the process for learning social network structure. In an online experiment, participants were asked to remember and recall the structure of a directed social network (i.e. a set of influence relationships among employees in a company). Participants were shown a list of names of nodes and asked to remember information about who (source) influences whom (target) in the network with a time limit. The information appeared on the display only when participants hovered the mouse pointer over a source node; in other words, participants needed to collect source-target information across all nodes in an efficient way. A social network stimulus was generated at each experimental session based on a random graph varying in three properties: size (5 to 20), density (.05 to .20), and the degree of reciprocity (0 to 1). Analysis of patterns of mouse tracking/fixation across nodes in a generalized linear mixed modeling (GLMM) framework revealed two fundamental strategies that significantly increased a recall rate of whole network structure. The first was a “chasing stars” strategy, or spending more time to remember active (more influential) nodes than inactive (less influential) odes. The second was a “confirming alliance” strategy, or searching pairs of nodes forming reciprocal relationships, and was notably efficient to improve the performance in the learning task of large social network structure. The current findings suggest that the exploration of information-searching processes for social network learning has potential to understand the nature of network schema/bias.