The human factors of interactive voice response: Designing for age inclusion.

摘要

Interactive voice response (IVR) was used to study the differences between younger and older adults in their use and acceptance of computer technology. A survey was completed by 139 people, and responses to attitude questions were compared to 2 previous studies done a decade ago. IVRs were not popular with callers then and much less liked now. The second part of the survey asked participants to rate applications of IVR as useful or annoying on a 7-point Likert scale. Several applications were considered useful, but most were considered annoying. In an experiment, 72 subjects (young = 40, old = 32) used a computer simulating an NR system to answer questions presented on cards. The data collected were response time and errors. The older group had slower response times and made more errors. Questions were formulated with 4 configurations of menus: 26, 43, 62 + 2, or 82. Two-choice menus (26) had six levels of menus to navigate. All the configurations had 64 targets except the 6-choice, which had 72. The 6-choice configuration had 2 menus of 6 choices plus a third menu with 2 choices. The fastest response times were with the 43 and 82 configurations. The number of errors monotonically decreased as the number of options per menu increased. Half of the test questions included menus with a suffix: an option at the end that the subjects were told to ignore. Subjects were able to ignore the suffix as evidenced by the lack of errors due to suffix, only slightly longer response times, and the fact that 79.2% said they could ignore it. Suffix was significant for response time but not for errors. Previous research seemed to suggest that the optimal number of options per menu is about 4 or 5 based on the capacity of short-term memory. Menus with more options and the ability to repeat and correct options may be preferable. Designers should strive to minimize error rather than maximizing speed since older adults are frustrated by committing errors in a computer application. Suggestions are made for further research employing usability testing and heuristic evaluation.