Does Users’ System Evaluation Influence Speech Behavior in HCI? – First Insights from the Engineering and Psychological Perspective


In interactions with speech based dialog systems users tend to adapt their speech behavior to their technical counterpart by taking care on the abilities and characteristics they ascribe to the system. Hence, it can be supposed, that different systems may evoke different speech behavior according to the users’ evaluation of the system. In order to support this hypothesis we compared a widely similar individualization-focused interaction between users and 1) a WOZ-simulated speech-dialog system (LAST MINUTE Corpus) and 2) a self-developed skill for Amazon’s Alexa. We examined measurable acoustic characteristics and subjective system evaluations operationalized by the item-based self-report questionnaire AttrakDiff. The analysis revealed that users’ speech behavior showed less acoustic variation in the interaction with Amazon’s Alexa. However, system evaluation did not differ significantly between the two experiments. In the discussion of this finding we, inter alia, take into account results from a first unsystematic analysis of interviews regarding users’ subjective experiences conducted after the interaction with Amazon’s Alexa.

Year: 2020
In session: Dialogue
Pages: 241 to 248