The dataset contains the data used to evaluate the designer’s empathic understanding during user interviews using the EA measure.
This data was collected from user interview sessions. The empathic accuracy dyadic interaction approach was used to carry out the empathic task. This approach evaluates the designer’s empathic ability with respect to the specific user in the interview. The process of data collection is decribed in the figure below:
In principle, empathic accuracy is the sentence pair similarity. However, in practice, it involves understanding the contextual meaning of sentences, which exceeds merely assessing text similarity.
The dataset consists of the timestamp, user’s thought or feeling, designer's guess, empathic accuracy scores, user’s emotional tone, designer’s guess of the user’s emotional tone and the emotional tone accuracy (ETA).
The description of the labels and the data collection process can be found in the publication below:
If you use or find this repository helpful, please cite :
Oluwatoba Fabunmi, Saman Halgamuge, Daniel Beck and Katja Holtta-Otto, 2024, "Large Language Models for Predicting Empathic Accuracy between a Designer and User".