Benefitting InfoVis with Visual Difficulties

• Published in: IEEE transactions on visualization and computer graphics Vol. 17; no. 12; pp. 2213 - 2222
• Main Authors: Hullman, J., Adar, E., Shah, P.
• Format: Journal Article
• Language: English
• Published: United States IEEE 01-12-2011
• Subjects: active processing; Cognition; cognitive efficiency; Comprehension; Computer Graphics; Data visualization; Desirable difficulites; engagement; Humans; individual differences; Learning; Models, Psychological; Psychology; Time factors; User-Computer Interface; Visual Perception
• ISSN: 1077-2626; 1941-0506
• DOI: 10.1109/TVCG.2011.175

Many well-cited theories for visualization design state that a visual representation should be optimized for quick and immediate interpretation by a user. Distracting elements like decorative "chartjunk" or extraneous information are avoided so as not to slow comprehension. Yet several recent studies in visualization research provide evidence that non-efficient visual elements may benefit comprehension and recall on the part of users. Similarly, findings from studies related to learning from visual displays in various subfields of psychology suggest that introducing cognitive difficulties to visualization interaction can improve a user's understanding of important information. In this paper, we synthesize empirical results from cross-disciplinary research on visual information representations, providing a counterpoint to efficiency-based design theory with guidelines that describe how visual difficulties can be introduced to benefit comprehension and recall. We identify conditions under which the application of visual difficulties is appropriate based on underlying factors in visualization interaction like active processing and engagement. We characterize effective graph design as a trade-off between efficiency and learning difficulties in order to provide Information Visualization (InfoVis) researchers and practitioners with a framework for organizing explorations of graphs for which comprehension and recall are crucial. We identify implications of this view for the design and evaluation of information visualizations.