Cursor Measures for Motion-Impaired Computer Users
Simeon Keates, Faustina Hwang, Patrick Langdon, P. John Clarkson, Peter Robinson · 2002 · Proceedings of the Fifth International ACM Conference on Assistive Technologies (Assets '02) · doi:10.1145/638249.638274
Summary
This paper systematically analyses cursor movement patterns of motion-impaired computer users to understand the specific difficulties they face with point-and-click interactions in graphical user interfaces. The authors argue that existing cursor measures, primarily developed by MacKenzie et al. for able-bodied users, focus on movement time and error rate as principal variables but treat cursor paths as "gross measures" — oversimplifying the rich behavioural data captured in cursor trajectories. The study collected cursor movement data from five motion-impaired participants (four with cerebral palsy and one with Friedreich's ataxia, including wheelchair users and one non-speaking user) and three able-bodied controls using a multidirectional point-and-click task based on ISO 9241-9. Seven existing MacKenzie measures were applied: target re-entry, task axis crossing, movement direction change, orthogonal direction change, movement error, movement offset, movement variability, plus movement time and throughput. The authors then propose six new cursor characteristics specifically designed for motion-impaired users: distance travelled relative to cursor displacement, distribution of distance over speed ranges, submovement count (based on a velocity threshold), cursor distance travelled away from the target, distribution of distance over curvature ranges, and distribution of distance over radii from the target.
Key findings
A two-factor repeated measures ANOVA confirmed a significant difference between able-bodied and motion-impaired users across cursor measures (F=13.33, df=1,287, p<0.001) and a significant interaction between impairment presence and cursor measure type (F=667.5, df=9,2583, p<0.001), demonstrating that cursor measures collectively capture meaningful differences between groups. Motion-impaired users showed notably higher values across most measures — for example, movement direction changes averaged 48.16 versus 13.28 for able-bodied users. A follow-up study using a Logitech Wingman force-feedback mouse with gravity wells around targets showed that haptic assistance significantly reduced several measures (F=40.77, df=1,599, p<0.001). However, only some measures — task axis crossing, orthogonal direction change, movement error, movement offset, and movement variability — were sensitive to the haptic assistance, while movement direction change, movement time, and throughput were not. The new cursor characteristics revealed distinct movement profiles: the most impaired participants (PI3, PI7) showed high-curvature movements, "hook backwards" patterns before reaching targets, and cursor distances greater than 700 pixels from target versus 287 for controls. Submovement analysis showed PI3 and PI7 had fewer but longer submovements, while able-bodied users made more brief submovements, suggesting fundamentally different motor control strategies.
Relevance
This paper makes an important methodological contribution to motor accessibility research by demonstrating that standard pointing metrics developed for able-bodied users are insufficient for understanding the difficulties of motion-impaired users. The six new cursor characteristics capture aspects of movement — path curvature, distance overshoot, submovements — that reveal why users struggle, not just that they do. For practitioners designing accessible interfaces, the findings have direct implications: gravity wells improve some aspects of cursor control but not all, and different types of motor impairment produce distinctly different cursor movement profiles, meaning that one-size-fits-all cursor assistance will not serve all users equally. The paper also demonstrates the practical challenges of research with motion-impaired participants — small sample sizes, heterogeneous impairments, and the need for long-term relationships with participants — challenges that continue to affect accessibility research validity. The proposed measures provide a framework for evaluating cursor assistance techniques that remains relevant as pointing interactions persist across desktop, tablet, and emerging spatial interfaces.
Tags: motor disability · pointing device · cursor control · target acquisition · force feedback · cerebral palsy · ataxia · haptic technology · user research · interaction design · Fitts's Law
Standards referenced: ISO 9241-9