Modeling a Graph Viewer’s Effort in Recognizing Messages Conveyed by Grouped Bar Charts Richard Burns 1 , Sandra Carberry 2 , and Stephanie Elzer Schwartz 3 1 Dept. of Computer Science, West Chester University, West Chester, PA 19383 USA rburns@wcupa.edu 2 Dept. of Computer Science, University of Delaware, Newark, DE 19716 USA carberry@cis.udel.edu 3 Dept. of Computer Science, Millersville University, Millserville, PA 17551 USA stephanie.schwartz@millersville.edu Abstract. Information graphics (bar charts, line graphs, etc.) in popu- lar media generally have a high-level message that they are intended to convey. These messages are seldom repeated in the document’s text yet contribute to understanding the overall document. The relative percep- tual effort required to recognize a particular message is a communicative signal that serves as a clue about whether that message is the one in- tended by the graph designer. This paper presents a model of relative effort by a viewer for recognizing different messages from grouped bar charts. The model is implemented within the ACT-R cognitive frame- work and has been validated by human subjects experiments. We also present a statistical analysis of the contribution of effort in recognizing the intended message of a grouped bar chart. 1 Introduction Information graphics are non-pictorial visual devices, such as simple bar charts, line graphs, pie charts, and grouped bar charts. They are incorporated into a mul- timodal document in order to achieve one or more communicative goals [12,11]. In the case of scientific documents, the communicative goal might be to present data or to help the reader visualize information. However, when information graphics appear in popular media such as periodicals (USA Today, Wall Street Journal) and magazines (The Economist, Time), they generally have a high-level message that they are intended to convey. For example, consider the graphics in Figures 1 and 2 which ostensibly convey that “Women are more likely than men to delay medical treatment” and that “food prices are lower in Iraq than in the United States” . Although the caption in Figure 1 explicitly states the graphic’s message, the caption in Figure 2 does not help recognize the message of that graphic. A study by Carberry et al. [5] found that a graphic’s message is of- ten not contained in the graphic’s caption or in the article accompanying the graphic. Yet the graphic’s message is integral to understanding the full content of a multimodal document. We are developing systems for recognizing the intended message of an infor- mation graphic in popular media. Our work has several applications. The first S. Carberry et al. (Eds.): UMAP 2013, LNCS 7899, pp. 114–126, 2013. c  Springer-Verlag Berlin Heidelberg 2013