While there has been significant research on how low-level perceptual elements contribute to a user’s ability to compare or discern data points, less is known about how seemingly meaningless properties of a visual scene contribute to the perception of information structure. We present the results of a study in which participants viewed five types of simple data visualizations that supposedly depicted information about the departmental structure of a series of companies. Although the underlying data was the same in each case, we altered simple design elements such as enclosure, connectedness of parts, whether parts were placed within a visible area, and whether the parts themselves were enclosed by borders. Participants were asked to rate each company on a series of semantic dimensions. The results show a significant effect of minor design elements on semantic interpretations of data, and comments by participants further suggest that these effects may be grounded in physical and emotional inferences derived from the appearance of charts.
Design Elements and the Perception of Information Structure. In IEEE Information Visualization Posters, 2009