Recent advances in computer graphics have led to the use of computerized representations of 3-D structures in fields such as engineering, architecture, science and medicine. While these 3-D images are projected on a 2-D screen, rotating the models helps to provide information about the depth and dimensions of the object. At each interaction with these 3-D images, an individual uses an external visualization, which is a visual-spatial representation viewed on a computer screen, to help with an internal computation, such as comprehension, learning, or planning.

For cognition researchers, Madeleine Keehner, et al., this interactive use of 3-D models raises the question, “If a person does not use an interactive visualization effectively, might a non-interactive visualization, showing them the most task-relevant information, be more effective?”

Their research, published in the January 2008 edition of the journal Cognitive Science, indicates that providing interactive control of computer visualizations may not necessarily improve an individual’s performance in spatial reasoning tasks. Rather, they found that viewing the most task-relevant information will enhance task performance regardless of whether the task-relevant information was obtained actively or passively.

In their study, the authors examined the effects of interactive visualizations and spatial abilities on a task requiring participants to infer and draw cross sections of a 3-D object.  They conducted three experiments that manipulated whether participants could interactively control a virtual 3-D visualization of the object while performing the task, and then the authors compared subjects who were allowed interactive control with those who did not have interactive control.

In the first experiment, the investigators found that subjects who were allowed to control the visual on the computer screen performed better, overall, than those who viewed a continuously rotating version of that model. However, in the second experiment, no differences were noted in the performance of subjects exposed to interactive and non-interactive images if the movements of the visualization were held constant in a “yoked” design. For this group, the quality of the visual information received was a more important predictor of performance, regardless of whether they controlled the image.

In the last experiment, a non-interactive condition was created to mimic the “ideal” movements of the visual, based on the most successful patterns of interactivity recorded in experiment two. The subjects did not have control of the visualization but were shown the most informative view on every trial. The outcome of this experiment revealed that these non-interactive participants performed better than the interactive subjects who could control model.

“As we continue to design interactive visualizations for cognitive performance, it is important to recognize that more interactivity may not always be better, manipulating external visualizations may not always be less effortful than relying on internal visualizations, and people may not always discover the affordances of external visualizations without explicit instruction,” the authors concluded.