Goal Crossing with Mice and Trackballs for People with Motor Impairments: Performance, Submovements, and Design Directions
Jacob O. Wobbrock, Krzysztof Z. Gajos · 2008 · ASSETS '08: Proceedings of the 10th International ACM SIGACCESS Conference on Computers and Accessibility · doi:10.1145/1361203.1361207
Summary
This study investigates whether goal crossing—an input technique where users move through a target boundary rather than clicking within it—can benefit people with motor impairments. The research compares crossing performance against traditional area pointing using a rigorous Fitts' law methodology with 16 participants: 8 able-bodied and 8 with motor impairments including Parkinson's disease, cerebral palsy, peripheral neuropathy, spinal cord injury, and spine degeneration. Participants performed reciprocal pointing and crossing tasks using both mice and trackballs at various target sizes and distances. The study employed comprehensive analysis including movement time, error rates, throughput (bits/s), path characteristics (direction changes, movement variability, movement error, movement offset), and detailed submovement analysis examining velocity, acceleration, and jerk profiles. The methodology addresses important questions about how motor impairments affect the fundamental mechanics of pointing versus crossing. While previous crossing research focused on able-bodied users, this study specifically examines whether crossing's reduced precision demands—requiring only that users pass through a boundary rather than stop within one—might benefit users who struggle with the fine motor control needed for conventional pointing and clicking.
Key findings
The most striking finding is that motor-impaired participants achieved 23% higher throughput with crossing (2.88 bits/s) than with pointing (2.34 bits/s), while able-bodied users showed the opposite pattern (4.72 bits/s pointing vs 3.61 bits/s crossing). This suggests crossing may be particularly advantageous for users with motor impairments. Path analysis revealed that motor-impaired users produced fewer direction changes and less movement variability during crossing compared to pointing. Perhaps most remarkably, jerk analysis showed that crossing nearly equalized motor smoothness between groups—the total squared integrated jerk was similar for both able-bodied and motor-impaired participants during crossing movements, while pointing showed dramatic differences. Error rates were higher for crossing (6.23% vs 1.94%), but analysis revealed many "errors" were actually crossings in the wrong direction rather than failures to acquire the target. Motor-impaired participants expressed strong preference for crossing, rating it as easier and faster than pointing.
Relevance
This research provides empirical evidence that alternative input techniques can significantly improve computer access for people with motor impairments. The finding that crossing equalizes motor smoothness between groups is particularly significant—it suggests that the right interaction technique can mitigate some effects of motor impairment rather than simply accommodating them. For accessibility practitioners, the study offers concrete design directions: two-stage crossing (crossing activates a widget, then another action commits), modal crossing (dedicated mode toggled by voice or eye gaze), and using submovement analysis to distinguish intentional crossings from incidental ones. The research also raises important questions about the "occlusion problem"—how to handle the fact that crossing requires a clear approach path to targets. While crossing-based interfaces remain uncommon in mainstream applications, this work demonstrates that fundamental interaction paradigms can be rethought to better serve users with motor impairments, potentially enabling faster and more comfortable computer use.
Tags: motor impairments · input devices · goal crossing · pointing · Fitts' law · submovements · trackballs · mice · HCI