A0586
Title: Using perceptual tomography for balance clustering in network construction and validation
Authors: Hsuan-wei Lee - Academia Sinica (Taiwan) [presenting]
Abstract: In social network analysis, when the information of socio-centric (i.e. whole) networks are difficult to get, researchers often use egocentric to understand the structure of network ties around an individual. In both methodological schemes, network sampling is widely used when self-reported ties are costly or could not be easily obtained. We combine network sampling approaches with third-party reporting: randomly chosen people are shown a random sample of photos from a group to which they belong and asked to group the photos according to whether or not the people in the photos are close to one another. Aggregated multiple 3rd-party data are analyzed to constructing the perceived presence or absence of individual properties and pairwise relationships, and a 3rd-party perceived network is thus built. This is a continuous work of the paper Mapping the structure of perceptions in helping networks of Alaska Natives. We generalize the parameters such as the number of clusters, the sampling size of the population, the number of photos that are shown to the sampled people and explore the theoretical implication of these change of settings. Lastly, we compare the perceptual networks and the classical self-reported social networks.
