Which Interface Makes Seniors Feel Safer?

A recent study examined how usage experience and input type (touch, speech, or eye control) affect trust and mental effort in older adults. Two groups participated: tech‑savvy seniors and those with little experience. After completing simple tasks with each modality, participants filled out the NASA‑TLX questionnaire to gauge cognitive load and a trust perception test. Two main trends emerged. First, experienced seniors showed higher trust in touch, relying on a well‑established “press‑and‑respond” mental model. Second, novices preferred speech, which mirrors natural conversation, cutting cognitive load by 24 % compared to touch. Yet, voice recognition errors quickly eroded their trust. Eye‑tracking data confirmed that speech demands less visual effort, but accuracy is critical. For designers, the message is clear: keep tactile interfaces for seasoned users, boost error tolerance for voice‑based systems aimed at beginners, and enhance visual feedback to lower tech anxiety. In this way, technology can become a more reliable ally for all seniors.

The study employed a 2 × 3 mixed experimental design with prior experience (experienced vs. inexperienced) and input modality (touch, speech, eye control) as factors. The primary cognitive load measure was NASA‑TLX, which aggregates six dimensions (mental demand, physical demand, temporal demand, performance, effort, frustration) into a weighted average score. Eye‑tracking data provided fixation metrics (duration, saccade amplitude) to assess visual effort during interaction. Two‑way ANOVA revealed a significant interaction between experience and input modality (p < 0.01) on perceived trust. Experienced seniors showed a +12‑point advantage on the trust scale with touch versus voice, while novices recorded a +15‑point advantage for voice. The average cognitive load for speech was 24 % lower than touch among novices (TLX mean 38 vs. 50). However, recognition error frequency (average 18 % failures) introduced a trust penalty, reducing it by roughly 10 points. From a human‑computer interaction perspective, these findings indicate that a “natural modality” (speech) is insufficient if recognition robustness is low. Cognitive load theory predicts that reducing extrinsic load (e.g., more intuitive interfaces) enhances learning, but only if errors do not add intrinsic load. Hence, for inexperienced users, implementing fault‑tolerance mechanisms—contextual feedback and confirmation prompts—is essential to sustain trust. Moreover, preserving tactile metaphors for experienced users supports cognitive continuity, minimizing the need to recalibrate mental models. In summary, designing interfaces for older adults must balance familiarity, recognition reliability, and operational transparency to optimize both trust and cognitive load.