SIGNIFICANCE: Personal banking activities are a crucial instrumental activity of daily living. Current accessibility standards for automated teller machines (ATMs) do not guarantee that all, or even most, users with disabilities can independently manage their banking needs. A prominent ATM manufacturer contracted with our center to compare the usability of a prototype design having enhanced accessibility features against an existing model.

INNOVATION: The methodology is based on universal design principles that aspire to make products as usable as possible for the greatest diversity of users. The approach was novel in four ways: (1) the research team included occupational therapists (OTs), architects, and human factors engineers; (2) four diverse user populations were studied; (3) three real-world interface conditions were evaluated using ATMs provided by a private sector sponsor; and (4) both observed and self-reported usability data were captured.

APPROACH: In this study, we attempted to answer the following research question: Is the usability of the prototype (ATM B) at least as good as the usability of the existing model (ATM A) for all participant groups?

Little research has been conducted on ATM usability for people with disabilities, and there is no industry standard design. The sponsor’s goal was to identify design changes necessary to ensure that ATM B would meet accessibility expectations for wheelchair users and people with vision impairment.

METHOD: This was a cross-sectional field study in which we evaluated the usability of two ATM designs with four groups. The test apparatus involved two machines: ATM A is a model currently sold to banks; ATM B is a prototype dual-screen design having a small screen and a large screen. Participants performed three typical transactions using each of three ATM screen conditions: ATM A, ATM B–large screen (ATM B–LS), and ATM B–small screen (ATM B–SS).

Four groups were recruited: wheelchair users (n = 11), adults with low vision who could use onscreen features (n = 10), adults with low visual who used the audio-output feature in lieu of the visual interface (n = 10), and able-bodied adults (n = 10). Participants were at least 18 yr old and were regular users of ATMs.

Participants rated the usability of each screen condition using three self-report measures: the Acceptability Rating Scale (ARS), the Difficulty Rating Scale (DRS), and the Rapid Assessment of Product Usability and Universal Design (RAPUUD). Qualitative data were gathered from participants using open-ended questions regarding each screen condition and staff field notes.

For each of the self-reported variables, a Kruskal–Wallis analysis of variance (ANOVA) was used to evaluate differences across the four participant groups and screen conditions.

RESULTS: Wheelchair users rated ATM B–LS significantly lower on the DRS (χ² = 18.36, p < .05), ARS (χ² = 14.56, p = .002), and RAPUUD Item 7 (χ² = 23.60, p < .05). Low vision users rated ATM B–SS significantly lower on the ARS (χ² = 17.33, p = .001), DRS (χ² = 20.54, p < .001), and RAPUUD Item 6: perceived assistance (χ² = 14.66, p = .002).

Anecdotally, wheelchair users preferred the small screen condition because it required a shorter reach distance, which increased their accuracy and success rates. Users with low vision preferred the large screen condition because of its larger screen size, larger font size, and better screen contrast.

CONCLUSION: The usability of the dual-screen prototype (ATM B) was acceptable for all groups.