Enhancing statistical chart accessibility for people with low vision: insights from a user test
Alcaraz-Martínez, Rubén,Ribera, Mireia,Pascual-Almenara, Afra
DOI: https://doi.org/10.1007/s10209-024-01111-4
2024-05-09
Universal Access in the Information Society
Abstract:A remote user test was performed with two versions (one accessible and one non-accessible) of three types of web-based charts (horizontal bar chart, vertical stacked bar chart, and line chart). The objectives of the test were: (a) to validate a set of heuristic indicators for the evaluation of the accessibility of statistical charts presented in a previous work (Fariñas Falcón et al. in Mediocentro Electrónica 21(1):65–68, 2017); (b) to identify new barriers and preferences for users with low vision in the access and use of this content not previously contemplated. 12 users were tested, with a variety of conditions associated with low vision: low visual acuity (6 users), reduced central vision (2 users), reduced peripheral vision (2 users), blurry vision (1 user), sensitivity to light (3 users), Nystagmus (2 users) and color vision deficiency (CVD) (4 users). From a quantitative standpoint, accessible versions of charts were more efficient, effective, and satisfactory. From a qualitative point of view, results verify the relevance of heuristics H2, Legend; H3, Axes; H6, Data source (as data table); H10, Safe colors; H11, Contrast; H12, Legibility; H13, Image quality; H14, Resize; H16, Focus visible; H17, Independent navigation; related to the proposed tasks. As new observations, tooltips were highly valued by all users, but their implementation must be improved to avoid covering up significant parts of the charts when displayed. The data table has also been frequently used by all users, especially in the non-accessible versions, allowing them to carry out tasks more efficiently. The position and size of the legend can be a significant barrier if it is too small or appears in an unusual position. Finally, despite the limitations related to color perception, some users prefer color graphics to black and white, so, to target all profiles, it is necessary to redundantly encode categories with colors and patterns as well.
computer science, cybernetics,ergonomics