A comparison of area pointing and goal crossing for people with and without motor impairments
Jacob O. Wobbrock, Krzysztof Z. Gajos · 2007 · Proceedings of the 9th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS '07) · doi:10.1145/1296843.1296847
Summary
Wobbrock and Gajos at the University of Washington present a controlled study comparing two fundamentally different target-acquisition paradigms — conventional area pointing (move into a confined region and click) and goal crossing (sweep the cursor across a goal line, no click required) — for users with and without motor impairments. The motivation comes from a body of prior work showing that motor-impaired users frequently slip out of small targets and that 28% of their mouse clicks contain movement during the click itself, and that specialized assistive devices are widely abandoned in favour of mainstream mice and trackballs. Sixteen participants — eight able-bodied and eight with motor impairments (cerebral palsy, spinal cord injury, Parkinson's disease, peripheral neuropathy, and spinal degeneration) — used a Logitech optical mouse, a Kensington Expert Mouse Pro trackball, or both, to perform a Fitts'-style task at five target widths (16–128 px) and three amplitudes (128–384 px) for a total of 6,240 trials. The authors fit Shannon-formulation Fitts' law models, computed throughput in bits/second, and ran path-analysis metrics (target re-entries, axis crossings, movement variability, error, and offset) developed by MacKenzie and extended by Keates. Mixed-effects ANOVA with Subject as a random effect was used to handle unbalanced device assignment within the motor-impaired group.
Key findings
Fitts' law fit both pointing and crossing for both subject groups (R² = 0.91–0.996), adding evidence to the disputed question of whether motor-impaired movement obeys Fitts' law. The headline result inverts between groups: able-bodied users had higher throughput for area pointing than for goal crossing (4.72 vs. 3.61 bits/s), but motor-impaired users had higher throughput for crossing than for pointing (2.88 vs. 2.34 bits/s). Crossing reduced movement time more for the motor-impaired group than for able-bodied users and was particularly advantageous on the trackball. Under a strict definition of crossing errors (passing the goal line beyond either end of the finite goal segment), error rates were higher for crossing than for pointing overall (6.23% vs. 1.94%), but the speed gain still dominated total acquisition time (MTε). Path analyses favoured crossing on most measures: less movement variability, less movement error, fewer perpendicular direction changes, and (for motor-impaired users) shorter total path distance. Velocity profiles indicate users adopt different strategies — ballistic-then-correct for pointing, smooth acceleration into a final "flick" for crossing. Subjectively, motor-impaired users preferred crossing and rated it easier and faster; able-bodied users preferred pointing.
Relevance
This paper is a foundational empirical case for designing accessible interfaces around crossing widgets rather than point-and-click, particularly for users whose impairments make it hard to land on and stay within a confined target while clicking. The argument is sharpened by the fact that the comparison runs on mainstream mice and trackballs — the devices motor-impaired users actually adopt — rather than specialised hardware. The practical takeaway is not that crossing should replace pointing universally, but that it should be available as an alternative interaction mode, especially in widgets where misses cost little (menus, scrubbers, scrollbars). Limitations the authors flag are honest and important: definitions of pointing and crossing errors are semantically different ("apples and oranges"); strict crossing errors inflate the error rate compared to what a real interface would penalize; small-target crossing (16 px) was especially error-prone for motor-impaired users; and the practical "occlusion problem" — one goal line obscuring another in mouse-based crossing — remains unsolved and is left for future work. For developers, this paper pairs naturally with Steady Clicks and the area-cursor literature as part of the toolkit for accessible pointing.
Tags: target acquisition · goal crossing · area pointing · Fitts' law · steering law · throughput · mouse · trackball · motor accessibility · motor impairment · cerebral palsy · spinal cord injury · Parkinson's disease · peripheral neuropathy · input methods · pointing device · cursor · accessible interfaces