Good symbol discrimimability (recognising that 2 symbols are the same or different) is vital for tactile map reading. This presentation discusses a study that was designed to produce a set of highly discriminable tactile symbols (points and lines) and to investigate what makes symbols discriminable from each other.
Forty point symbols and 20 line symbols were selected for this study, based on previous research and existing use. Thirty visually impaired and thirty blindfolded sighted participants were asked to test these symbols in a paired-comparison, same-different judgement task. Discriminability was measured by the number of confusions; the smaller the number of confusions between a given pair of symbols, the more discriminable they were. This methodology allowed us to produce a highly discriminable set of tactile symbols, consisting of 11 point symbols and 4 line symbols.
In addition to producing a set of discriminable symbols, we explored the factors that underlie symbols discriminability. Using multi-dimensional scaling, the discriminability of symbols was represented in a two dimensional graphic, where larger distances between symbols indicated higher discriminability and similar symbols clustered together. These plots allowed us to categorise symbols and speculate on the bases for tactile symbol discriminability. Using these generalisations, it might be possible to generate further symbol sets that are highly discriminable.
Prior knowledge of topic assumed: Low
Contains advanced technical content: No
Back to the programme list
Back to the 2005 conference home page