This DDfV seed project adresses the epistemically responsible design of explanation-aware algorithms and interfaces. Two proof-of-concept studies will show how perspective and affect influence critical thinking and motivated numeracy.
This project addresses ethics and values in data science, with the aim of designing epistemically responsible debate structures. The researchers investigate how controversial information is debated online, and how technology can be used to support the epistemic and dialogical aspects of such communication. This project aims to answer how affect, emotion, and sentiment influence debate about contentious topics. In so doing, the project will enhance and enrich the design of explanation-aware algorithms and explanation-aware interfaces that consider human factors when helping people make decisions together with artificial advice givers, as detailed in the figure below.
To reach these aims two proof-of-concept studies will be conducted:
- Analysis of critical thinking in online debate. The researchers compared a reply- and retweet-network on the topic of vaccine safety from Twitter with a debate network retrieved from Rbutr. This analysis compared indices of the quality of debate, such as sentiment and grammatical sophistication, in these two corpora. Differences between the two platforms were found.
- Selective numeracy. A large (N=1600) experiment with end-users to study how emotionally-charged debate about contentious topics like immigration can engage or disengage numeracy and other critical thinking capacities. This is a reproduction study, and the PIs have collaborated with colleagues at UC Berkley (USA) to conduct the necessary analysis.
Researchers in this project