A Comparison of Touchscreen and Mouse for Real-World and Abstract Tasks with Older Adults
Sung-Hee Kim, Kailun Zhang, Joanna McGrenere, Kellogg S. Booth, Claudia Jacova · 2020 · ACM Transactions on Accessible Computing · doi:10.1145/3418057
Summary
This study investigates whether input device choice (touchscreen vs. mouse) affects performance on computerized cognitive tests for older adults, with implications for the C-TOC (Cognitive Testing on Computer) project developing self-administered online cognitive screening. As older adults increasingly take cognitive tests remotely using whatever device they have at home, understanding device-related performance differences becomes critical for valid assessment. Sixteen older adults (ages 57-88, M=71.81) completed four C-TOC cognitive subtests—Picture-Word Pairs (memory/low-precision pointing), Arithmetic (numeracy/high-precision pointing), Sentence Comprehension (working memory/low-precision dragging), and Pattern Construction (visuospatial/high-precision dragging)—along with corresponding abstract Fitts's Law tasks on both an iPad touchscreen and a MacBook with mouse. The researchers examined whether Fitts's Law performance differences between devices could predict cognitive task differences, and analyzed movement patterns during the Pattern Construction dragging task to identify qualitative behavioral differences between devices.
Key findings
Touchscreen was faster than mouse for three of four cognitive tasks: 32% faster for Picture-Word Pairs, marginally faster (12%) for Arithmetic, and significantly faster for Sentence Comprehension. However, accuracy was significantly affected only on the high-precision Arithmetic task, where mouse users achieved 90% accuracy versus 82.5% for touchscreen. Critically, Fitts's Law task performance could not reliably predict device differences on cognitive tasks—correlations explained less than 50% of variance, insufficient for automatic calibration. Movement pattern analysis revealed striking behavioral differences: on touchscreen, participants made nearly 50% more moves (24 vs. 15 per trial) but with shorter distances (50px vs. 106px), decomposing single dragging movements into multiple "sub-moves." This sub-move strategy appears driven by finger occlusion (blocking view of target), touchscreen friction, and fail-to-grab errors unique to touch input. Despite these differences, total task completion time for high-precision dragging was equivalent across devices—the sub-move overhead on touchscreen was offset by faster individual movements.
Relevance
This research provides essential guidance for designing accessible interfaces for older adults and for developing valid computerized cognitive assessments. The finding that device choice affects speed and accuracy differently across task types has direct implications: interfaces for older users should favor pointing over dragging on touchscreens, and high-precision tasks require larger touch targets (minimum 40px, ideally larger) to avoid the accuracy penalty observed with smaller targets. The sub-move behavior suggests touchscreen interfaces should support "decomposable dragging"—allowing users to drop and re-acquire objects mid-drag without penalty. For healthcare applications, the inability to predict cognitive task performance from Fitts's Law tasks means device-specific normative data will be required; a score obtained via touchscreen cannot be directly compared to one obtained via mouse. The study also highlights a training gap: many older adults still treat capacitive touchscreens like resistive ones, pressing hard or using fingernails, suggesting explicit instruction improves interaction quality.
Tags: older adults · aging · touchscreen · input devices · cognitive testing · motor control · Fitts Law · pointing · dragging
Standards referenced: ISO 9241-400