The influence of word-length vocabulary variation in chatbots on the user’s attribution of mind to the chatbot, in relation to the user satisfaction.
Keywords
Loading...
Authors
Issue Date
2022-01-30
Language
en
Document type
Journal Title
Journal ISSN
Volume Title
Publisher
Title
ISSN
Volume
Issue
Startpage
Endpage
DOI
Abstract
Previous studies indicate that while chatbot research has come a long way technologically, the social aspect of chatbot research needs to be attended to. More specifically: how much ‘mind’ the end user attributes to the chatbot. A reliable and objective method to research the user’s attribution of mind to chatbots is to observe the user’s behaviour, given some variation of a specific variable. In this thesis, the word-length is varied, as it has been shown that this variable influences human behaviour. Therefore, this study’s purpose is to examine whether a person’s mind perception gets influenced by variation of word-length in chatbot vocabulary.
Subsequently, this study examines whether differences in mind perception influence the user satisfaction. A study by Bartneck et al. (2008) has shown that one of the overarching goals of chatbot research is to maximize user satisfaction, as this defines the performance criteria of chatbots.
To investigate the mind perception with such variation, an educational chatbot quizzing users on internet security was created. Subsequently, the subjects converse with one of three versions of the chatbot: high frequency of short words, high frequency of long words and a mix of the former and latter. Then a human-robot interaction surveys, adapted from Bartneck et al. (2008) and Balaji, D., & Borsci, S. (2019) are filled out and analysed.
The findings suggest that there is no such influence present of the word-length variation on the mind attribution by end users. The findings also suggest that the mind attribution by end users and the user satisfaction are highly correlated. This means that the more mind the end user attributes to the chatbot, the higher the user satisfaction will be.
Description
Citation
Faculty
Faculteit der Sociale Wetenschappen