Opinion prediction as communicative success, and the role of visual information in conversations
Author
Summary, in English
Abstract in Undetermined
We devised a measure based on predictions of a conversation partner's opinion in order to approximate
communicative success in an unrestricted conversation with or without access to visual information.
We validated this prediction-based measure against our intuitions that participants would have lower
prediction errors in conversations where the participants rated that they contributed more and were
more engaged in. Our results show that visual information in the form of images accompanying the
conversation topics decreased the prediction errors of the participants, but only for images that were
rated as helpful.
We devised a measure based on predictions of a conversation partner's opinion in order to approximate
communicative success in an unrestricted conversation with or without access to visual information.
We validated this prediction-based measure against our intuitions that participants would have lower
prediction errors in conversations where the participants rated that they contributed more and were
more engaged in. Our results show that visual information in the form of images accompanying the
conversation topics decreased the prediction errors of the participants, but only for images that were
rated as helpful.
Department/s
Publishing year
2012
Language
English
Publication/Series
[submitted]
Document type
Journal article
Topic
- Human Aspects of ICT
Keywords
- visual world
- pragmatics
- visual context
- mind-reading
Status
Unpublished
Project
- Thinking in Time: Cognition, Communication and Learning