Developing Voices for Smart Homes: User Personality and Synthetic Voice Interaction

Abstract

Participants (n = 40) aged 27 through 76 years (M = 57.7, Mdn = 59.0) took part in a 2 (voice emotion) x 2 (message content) within subjects factorial design experiment with the between-subject covariates agreeableness and emotional stability. Additional exploratory analysis included age and gender as between-subject factors. Analysis indicated that voice emotion level does not interact with message content level to determine user-reported message suitability ratings, with or without age and gender. No known experiments have been conducted (a) concerning the simultaneous examination of our three experimental factors, (b) in this population of seniors, and (c) through the World Wide Web. We intended, therefore, with this research design to make a unique contribution to HCI and speech technology.