Joint Analysis of Static and Dynamic Importance in the Eye-Tracking Records of Web Page Readers

Abstract

Heat maps highlight cumulative, static importance in eye-tracking records, while network analysis helps to elucidate dynamic importance from transitional relations. The present study was designed to perform both analyses in the same conceptual framework, i.e., network representation. For this purpose, heat maps comprising 5 × 5 segments were overlaid with networks, both of which were produced from the eye-tracking records of 20 subjects who read 10 top web pages that were classified into three layout types. The heat of the segments was graded on the basis of five percentile scores, whereas the core-peripheral nodes were identified by the agreement of centrality and ranking indices. The congruence between the two types of importance was generally good at the node level and the community levels. Additional findings included a) mixed patterns of the sustained fixations (i.e., loops) within the total fixations, and b) an increase in reciprocity as the network scope was narrowed to communities and then to the core neighborhoods.