Can voice interfaces assist experimenters in collecting data for research studies?

Description

When it comes to voice assistants (VA) like Siri or Alexa, little research has looked at how their conversational styles affect older adults with different personalities. In this study, 34 middle-aged adults (ages 50 to 64) and 24 older adults (ages 65 to 80) interacted with RAVA, a Wizard of Oz simulated VA acting a “research assistant” (in a simulated home setting using either formal or informal language. The results showed that RAVA was successfully able to collect data and older adults with higher levels of agreeableness thought the VA was more intelligent and likeable than middle-aged adults did. Additionally, older adults preferred informal VAs, especially those who scored low on agreeableness. Middle-aged adults, on the other hand, had similar acceptance levels for both formal and informal VAs. Interestingly, both middle-aged and older adults often thought of VAs as being like humans, but older adults used different metaphors depending on the VA's conversational style. Formal VAs were compared to professionals like librarians or teachers, while informal VAs were compared to close family members like spouses or relatives. Overall, this study shows that the conversational style of VAs can have different effects on middle-aged and older adults, highlighting the importance of personalized design.