Interaction Design and the Role of Spatial Ability in Moderating Virtual Molecule Manipulation Performance

Interaction Design and the Role of Spatial Ability in Moderating Virtual Molecule Manipulation Performance Trevor J. Barrett (trevor.barrett@psych.ucsb.edu) Department of Psychological and Brain Sciences, UCSB Santa Barbara, CA 93106 USA Mary Hegarty (mary.hegarty@psych.ucsb.edu) Department of Psychological and Brain Sciences, UCSB Santa Barbara, CA 93106 USA orientation and configuration of 3D molecular models. The literature suggests mixed effects of stereo viewing in virtual environments. Some studies report significant performance benefits from providing stereo viewing (Wang, MacKenzie, Summers, & Booth, 1998; Arsenault & Ware, 2004); while other studies report null effects of providing stereo (Khooshabeh & Hegarty, 2010; Barrett & Hegarty, 2013). The literature regarding colocation of haptic and visual workspaces is similarly inconsistent, with some studies reporting significant performance benefits (Ware & Rose, 1999; Barrett & Hegarty, 2013), and others reporting null effects (van Liere, Martens, Kok & van Tienen, 2005). Significant interactions between display and interface technologies have also been reported (Ragan, Kopper, Schuchardt & Bowman, 2012). The heterogeneity of findings in the virtual interaction literature likely arises from differences in experimental task demands. For example, performance benefits from providing stereo viewing are likely to be observed only when the added fidelity of the third dimension provides relevant or necessary information required for the particular task. Thus, we should be cautious in generalizing the effects of display and interface technologies, as their merit depends on the specific characteristics of the task at hand. Virtual technologies may also be differentially effective for people of different spatial abilities. While it is widely accepted that individual differences in spatial ability play an important role in learning from 3D virtual environments, the nature of aptitude-treatment-interactions (ATI) are widely disputed in the literature (Hoffler, 2010). Two main ATI hypotheses have been offered in studies of multimedia learning, animation, and interactivity. The ability-as- enhancer hypothesis predicts that high ability individuals are uniquely able to utilize increased fidelity to improve performance, whereas low ability individuals do not profit due to cognitive overload. The ability-as-compensator hypothesis predicts that individuals with high ability are able to compensate for lower fidelity representations and do not benefit from increased fidelity, whereas individuals with low ability benefit from increased fidelity because rich external representations can supplant or compensate for lack of ability. Of course these two hypotheses are not mutually Abstract Virtual models are increasingly employed in STEM education to foster learning about spatial phenomena. However, the role of design and spatial ability in moderating performance are not yet well understood. We examined the effects of display fidelity (stereo vs. mono), interface location (colocated vs. displaced), and spatial ability on performance during a virtual molecule manipulation task. The results indicated a significant beneficial effect of providing stereo viewing on response time, while interface location had no effect. The effect of providing stereo on performance was moderated by spatial ability. Notably, providing stereo did not benefit higher spatial ability participants, while those with lower spatial ability uniquely benefited from using the higher fidelity stereo display. Keywords: spatial cognition; individual differences; stereo; colocation; display; interface; virtual; rotation Computer-based virtual models are now an important instructional medium in science, technology, engineering, and mathematics (STEM) education (Trindade, Fiolhais & Almeida, 2002). Three-dimensional (3D) virtual models have shown promise in fostering meaningful learning; however, the perceptual cues provided often vary from system to system, leaving much to be understood regarding the impact of interaction design on reasoning and learning. Effects of virtual model design elements depend greatly on the given task as well as individual ability level. Given the increasing availability of new display and interaction technologies, building generalizable theories will be essential in understanding how to best design cognitively supportive virtual environments for education and research. Here, we examine the effects of display fidelity (stereoscopic vs monoscopic viewing), interface location (collocated vs displaced), and spatial ability on a virtual object manipulation task. Regarding display fidelity, stereo displays use binocular disparity to create the illusion of depth when viewing 3D content, whereas traditional displays provide only monocular depth cues. Regarding interface location, motion tracked interfaces for manipulating virtual objects may be displaced from the virtual image (e.g. typical computer mouse position) or colocated with the virtual object to minimize disparity between visual and haptic information. In the virtual object rotation task studied in our experiment, students used a direct manipulation interface to match the