Abstract
Gathering and integrating information from spatially distinct and qualitatively different sources requires decisions about when and where to sample information. An example is reading graphs and accompanying text to arrive at a coherent interpretation. Observers inspected graphs depicting realistic data, along with associated text, in one of the following configurations: (1) Simultaneous: Graph and text were adjacent, requiring a saccade to switch between them; (2) Eye contingent: Graph and text appeared sequentially, with the appearance of each triggered by a saccade, and (3) Button-press: graph and text appeared sequentially, with the appearance of each triggered by a button press. Observers answered a true/false question based on the graph after 60s of viewing (response accuracy~70%). About twice as much time was spent reading the text as inspecting the graph, despite the greater utility of the graph in answering the t/f question, and even when text content was tangential to the graph. During the initial reading of text, Ss rarely stopped mid-passage to switch to the graph. The perceptual configuration and the motor action needed to switch between graph and text had surprisingly large effects. The average number of transitions/trial between graph and text was greater in the simultaneous (6.5 ± 0.5) than in either the eye-contingent (4.5 ± 0.5) or button-press conditions (3.3 ± 0.5). Corresponding mean durations of individual visits to graph or text were 7 s for simultaneous, 8.5 s for eye-contingent and 12 s for button-press conditions. These results show that strategies used to sample information from spatially-distinct, heterogeneous sources depend not only on the information value, but also on perceptual availability and motor effort. Preferences to avoid even modest increases in perceptual-motor cost may be useful in order to preserve limited resources for processing the content and meaning of the displays.
Meeting abstract presented at VSS 2016