Talk-based versus text-based: the effects of wiki-surveys on attitudinal change in a deliberative mini-public

This package presents data from a study comparing two different approaches to online deliberative mini publics - one talk-based and the other text-based. The study invited 30 representatively sampled members of the British public to deliberate on issues associated with response personalisation in artificial intelligence chatbots following completion of a learning component. Talk-based deliberators gathered for 50 minutes over video conferencing to have small-group, semi-structured discussion with human facilitation. Text-based deliberators used the site AllOurIdeas to complete a 'wikisurvey' where participants responded to an open-ended prompt by writing responses and voting on pairs of responses submitted by others. Several dimensions of deliberative engagement and participant attitudes were measured as outcome variables through pre and post-deliberation surveys. The raw data presented includes responses to all survey questions, as well as discussion transcripts and submissions produced during the text-based exercise. The accompanying draft paper provides further context.