Is Speech Recognition Technology Robust for Older Populations?

Abstract

The primary objectives of this study were to investigate the robustness of current speech recognition software for older people and to gather data on the acceptability of this technology for this population. The sample included twenty-four adults from two age groups: young (20 - 29 yrs.) and older (66 - 80 yrs.). The speech recognition software used for the study was a commercially available continuous speech recognizer. Participants used the software to input a passage of text for five trials. Performance measures included reading time and system recognition errors. The results indicated that the reading time of the older adults was significantly longer than that of the younger adults. However, despite this difference there were no age differences in system recognition errors. In addition, the older people found the system easy to use and were comfortable using the technology. These results suggest that current speech recognition software is robust for both younger and older adults and is useable and acceptable by older people.