Partial overlap and local misalignment between visual attention, aesthetic preference, and expert-assessed cultural value

Published: 6 November 2026| Version 1 | DOI: 10.17632/fvs2xxb2pd.1
Contributor:
Haowei Wang

Description

Yunnan–Tibet wooden watchtower dwellings are conserved component by component, yet evidence for priority is qualitative. Visual attention, aesthetic preference, and expert-assessed cultural attributes were measured on 24 door-component forms drawn from 100 field samples, using eye tracking with 34 viewers (ten metrics reduced by PCA), ratings by 36 raters, and two Delphi rounds with ten experts (Kendall’s W = 0.700). Global visual activity was associated with preference (r = 0.608), whereas a fixation-allocation contrast was not (r = −0.020). Ethnic fusion was the only facet associated with preference after Holm correction (r = 0.634, p = 0.009) and was nominally associated with attention (r = 0.482, nominal p = 0.017; Holm-adjusted p = 0.136). Together, VA1 and EA2 explained 52.1% of the variance in-sample (leave-one-out R² = 0.344). Kendall τ among the three dimensions ranged from 0.36 to 0.51, and 9 of 24 forms differed by six or more ranks between visual attention and expert-assessed ethnic fusion, indicating local rather than systemic overlap. The ethnic-fusion association rested on a few high-scoring forms, and visual attention on one case (D4-1). Line drawings retain form but remove color, pattern, and use context, so the ethnic-fusion-only result is a boundary condition of the stimuli, not evidence that the facets are independent. Attention, preference, and expert value overlap only where cultural meaning is visually legible under this operationalization, which provides a basis for component-level screening. The framework thus juxtaposes expert judgment and viewer response at the component level without collapsing either into a single index, offering comparable evidence for conservation screening.

Files

Categories

Consumers' Aesthetic Response, Visual Culture

Licence