Abstract
People can interpret abstract messages encoded in colors, making color a useful feature for visual communication. Evidence suggests that when people interpret color-coding systems, they infer mappings that maximize the total association strength between colors and their assigned concepts; the assignment hypothesis (Schloss, Lessard, Walmsley, & Foley, 2017). The assignment hypothesis implies that people's interpretations of color-coding systems will vary depending on the degree to which different concepts are activated in their minds while solving color-concept assignment problems. We tested this prediction using a recycling paradigm. Participants saw pairs of different colored bins (unlabeled) and determined which bin was appropriate for discarding different kinds of objects (paper/trash). All participants had at least one kind of object to discard (paper/trash), so that object's concept was highly activated in their minds during the experiment. Across participants, we varied the degree to which a second object was activated. The "one-one" group was told about one object and discarded one object, so the second object was not activated. The "two-one" group was told about both objects, but discarded one object, so the second object was moderately activated. The "two-two" group was told about and discarded both objects, so the second object was strongly activated. We constructed a model to predict how responses would differ across groups if all participants solved an assignment problem to interpret the color-coding system on each trial, but the groups differed in the weight given to the second object (one-one group: weight=0; two-one group: weight=0.5; two-two group: weight=1). The model strongly predicted variations in color selections across groups (r=0.85, p < .001), suggesting people infer optimal color-concept assignments depending on the input they have. The results emphasize the importance of considering both the colors in the scene and the activated concepts in people's minds for designing intuitive color-coding systems.
Meeting abstract presented at VSS 2018