Imulating storyteller-audience interactions in digital storytelling: questions, exchange structures & story objects

Abstract

This work revolves around the design and evaluation of digital storytelling simulating real personal storytelling. Study One was an ethnography, of real storytellers, which revealed types of narratives, dynamism and interactivity in storytelling. This was used to design digital storytelling, which simulated the behaviours of real storytellers. Three design ideas, questions, exchange structures and story objects , were prototyped and evaluated in Studies Two, Three and Four. Study One took place over three months at the District Six Museum, Cape Town. We studied narratives from three guides about their Apartheid-era experiences. Discourse analyses showed the narratives: (a) were structured as clauses, each relating a story event or thought; (b) varied minimally across retellings; (c) incorporated storyteller-audience interactions (periodic questions) between clauses which matched teacher-student interactions described by Sinclair & Coulthard (1975); and, in exchange structures, guides periodically asked audiences questions; and (d) incorporated the museum exhibit and memory box objects. The digital storytelling design focused on: simulating questions and exchange structures; and story objects, allowing user-triggered narratives. We implemented a virtual environment containing two interactive storyteller agents, and several story objects. Study Two (n=101) manipulated the effect of questions and exchange structures on story experience. Study Three (n=69) manipulated the effect of story objects on story experience. Story experience was composed of: interest in the narrative context, enjoyment of and engagement in the storytelling, and the storytelling realism. These were measured with a questionnaire created for these studies; psychometric analysis showed it to be valid and reliable. Linear models showed questions increased interest (F =5.72, p =0.02) and engagement (F= 3.92, p =0.05) while exchange structures increased interest ( F =6, p =0.02), enjoyment ( F =4.14, p <0.04) and engagement ( F =10.53, p =0.002). Usage logs showed participants interacted readily with both while the agents could answer a mean of 35% of user questions. Story objects did not impact story experience. Study Two and Three's participants reported high story experience scores and predominantly positive qualitative feedback. In Study Four (n=93), the prototype was exhibited at District Six Museum for nine days. We observed visitor interaction, logged usage automatically and gathered voluntary feedback, which was largely positive. Visitors tended to engage passively with the prototype and linear models showed age was a predictor of the number of question ( F= 31.75, p <0.001) and exchange structure ( F =4.45, p <0.04) inputs. Additionally, multiple visitors would use the prototype simultaneously. We conclude that integrating different methodologies allowed us to simulate real storyteller-audience interactions and that the questions and exchange structure interactions we designed improved experiences of digital personal narratives. This design may be replicated by others seeking to similarly preserve the experience of personal storytelling.