Mixed Graph Designs Do Not Improve Visual Memory

Objects that are distinct in their categories and features tend to be better for visual memory, this is thought to be due to a decreased amount of overlap for object representations and interference (Konkle et al., 2010). If this is the case, this distinct advantage may have implications for improving memory in real-world situations, such as trying to remember multiple patterns of visualized data shown across the sequence of a research presentation. Previous research on individual visualizations has shown that visualized data is more memorable when visually distinct (e.g. contains a drawing) compared to plain graphs (Bateman et al., 2010). However, no existing work has tested whether this advantage in visual distinctiveness can be applied to the data shown across a series of graphs. This study asked 144 participants to recall specific trends and higher-level patterns (i.e. overarching trends) in data within a realistic journalistic data story. The graphs were either of distinct designs (e.g., bar, pie, and line graphs in one series) or identical designs (e.g. three bar graphs in one series). We find that distinctiveness in graphs had no statistically reliable effect on recall accuracy, suggesting that mixed graphical designs do not improve memory.