University of Twente Student Theses
Exploring lexical alignment for understandable information from trusted healthcare chatbots
Schaaij, Keara (2024) Exploring lexical alignment for understandable information from trusted healthcare chatbots.
|
PDF
8MB |
| Abstract: | This research investigates the impact of lexical alignment in a healthcare chatbot on the user’s understanding of information provided by the chatbot and their trust in this chatbot during an information-seeking task. Alignment has been shown to enhance human-agent dialogue, therefore, lexical alignment is explored to address discrepancies in language use between healthcare professionals and patients due to varying health literacy levels (referring to the ability to obtain, understand, and use health information and services to make appropriate decisions regarding one’s health). Therefore, a healthcare chatbot was developed that either lexically aligned with the user or did not. An experiment was conducted with 24 participants randomly assigned to one of two conditions (n=12 per condition). Alignment was achieved by substituting placeholders in the chatbot’s responses with participants’ preferred terms. The findings showed no statistically significant difference in understanding or trust due to lexical alignment. However, alignment decreased the perceived difficulty of the chatbot’s language and positively influenced clarification-seeking behaviour. Moreover, trust was primarily attributed to the perceived trustworthiness of the sources used by the chatbot. In conclusion, this research revealed that lexical alignment positively influenced interaction and language perception, with the main contributors to trust being the trustworthiness of sources. |
| Item Type: | Essay (Master) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 18 languages and literature, 54 computer science |
| Programme: | Interaction Technology MSc (60030) |
| Link to this item: | https://purl.utwente.nl/essays/102825 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page