Leveraging large data sets for user requirements analysis

In this paper, we show how a large demographic data set that includes only high-level information about health and disability can be used to specify user requirements for people with specific needs and impairments. As a case study, we consider adapting spoken dialogue systems (SDS) to the needs of older adults. Such interfaces are becoming increasingly prevalent in telecare and home care, where they will often be used by older adults. As our data set, we chose the English Longitudinal Survey of Ageing (ELSA), a large representative survey of the health, wellbeing, and socioeconomic status of English older adults. In an inclusion audit, we show that one in four older people surveyed by ELSA might benefit from SDS due to problems with dexterity, mobility, vision, or literacy. Next, we examine the technology that is available to our target users (technology audit) and estimate factors that might prevent older people from using SDS (exclusion audit). We conclude that while SDS are ideal for solutions that are delivered on the near ubiquitous landlines, they need to be accessible for people with mild to moderate hearing problems, and thus multimodal solutions should be based on the television, a technology even more widespread than landlines.