Running head: Inspiring Stories through Content Inspiring Stories through Content

Through user testing of Storytelling Alice, we found that users’ ability to find and develop story ideas was important in maintaining their engagement with programming in the system. In this paper, we describe the design process behind creating a gallery of characters and scenery that helps Storytelling Alice users to find and develop story ideas as well as lessons learned about successful and unsuccessful strategies for scaffolding the process of finding story ideas through content. We analyze thirty-six stories created with Storytelling Alice and examine the relationship between story complexity, users’ attitudes towards Storytelling Alice, and their programming behavior. Inspiring Stories through Content 3 Storytelling Alice is a programming environment designed to make the process of learning to program more appealing to middle school aged girls by focusing on the activity of storytelling (Kelleher, 2007). A recent study comparing girls’ programming behavior using Storytelling Alice and a version of Alice without storytelling support (Generic Alice) found that girls who used Storytelling Alice spent 42% more time within the system programming (as opposed to doing non-programming tasks such as using the mouse to position objects in the 3D scene) (Kelleher, 2007). Further, where only 17% of Generic Alice users snuck extra time to program, 51% of Storytelling Alice users snuck extra time (Kelleher, 2007). In early user testing of Storytelling Alice, we noticed that users’ ability to find a story idea they were interested in pursuing seemed to be an important in maintaining their engagement in the process of programming. Further, the characters and scenery that girls added to their Alice worlds often had a substantial impact on their ability to find a story idea, their success in creating a program, and on their continuing interest in using Storytelling Alice. The potential impact of girls’ choices of 3D objects was illustrated by a pair of girls who came in to user test an early version of Storytelling Alice. One of the two girls chose to add a dinosaur and a person to her world. She then proceeded to build a simple story in which the dinosaur scared the person and the person ran away in fear. Having accomplished that, she added a mouse character and continued her story by having the dinosaur be frightened of the mouse and run away. In this case, the dinosaur’s potential to be frightening provided inspiration for a simple story. Inspiring Stories through Content 4 The other girl was drawn to a collection of amusement park models, in part because it was one of the only cohesive spaces available in the gallery at the time. She spent a long time carefully arranging the rides in her amusement park and then added a man into the park. She began by having the man ride the merry-go-round but quickly ran out of ideas she wanted to pursue. Lacking a goal she wanted to pursue, she quickly lost interest in interacting with the system. In response to our observations that the content users chose for their stories could contribute to their success or failure at finding a story, we began to examine how to design content for a storytelling system that helps users find and develop ideas. In this paper, we describe the design process behind creating a gallery of characters and scenery for Storytelling Alice as well as lessons learned about successful and unsuccessful strategies for scaffolding the process of finding story ideas through content. Based on an analysis of thirty-six stories produced with Storytelling Alice, we found that story complexity correlates strongly with the percentage of time users spent within Storytelling Alice on programming tasks (rather than 3D scene layout tasks). Related Work There has been extensive past research on creating storytelling systems for children. The goals of these systems range from entertainment to improving literacy to enabling selfexpression. One family of storytelling systems helps users to improve their literacy skills. StoryMat (Cassell, 2004) and Rosebud (Cassell, 2004) encourage children to tell stories and, based on their stories, play back related stories. User studies of StoryMat have found that children incorporate Inspiring Stories through Content 5 elements of stories they hear into their own stories (Cassell, 2001). Sam (Cassell, 2004) is a virtual peer who listens to a child’s story and replies with a more complex story that models language skills the child has not yet mastered. SAGE allows children to create virtual storytellers, a task which requires that they both listen to and tell stories. Using TellTale (Glos, 1997) and FamilyBlocks (Glos, 1997) children can rearrange stories and pieces of stories with a set of physical props, enabling the exploration of different story progressions. Through encouraging children to reflect on their own and others’ stories, these systems help children to become better storytellers. Researchers at the University of Maryland working in inter-generational design teams have created several storytelling technologies. KidPad (Druin, 1997) is a spatial storytelling system in which children can draw a series of pictures and connect them together using hyperlinks in a zoomable interface. StoryRooms (Alborzi, 2000) are interactive story spaces in which the actions of Story Room visitors can gradually reveal a story. In the process of creating StoryRooms, the authors found that props could be helpful in getting an inter-generational team to tell a collaborative story. This observation lead to the creation of “idea cards” which can depict an object, setting, or an event and are intended to inspire story ideas (Alborzi, 2000) in the StoryRoom space. To enable children to create their own StoryRooms, researchers have prototyped a rule-based programming system (Montemayor, 2001). Commercially available programs such as American Girl Premiere (Learning Company) and Barbie StoryMaker (Mattel) allow children to create short animated movies using pre-made animations. Programming systems that enable users to create animations such as ToonTalk (Kahn, 1996), EToys (Kay), Scratch (Maloney, 2004), and StageCast Creator (Smith, 1994) can Inspiring Stories through Content 6 be used to create animated stories, but do not contain explicit support for finding or developing story ideas. While our work was inspired by an observation similar to the StoryRooms observation that props can help users to find story ideas, our goal in this research was to develop an understanding of how the attributes and presentation of content can influence users’ success telling stories. Formative Testing – Story Kits To begin investigating how content can help users generate ideas for their stories we decided to explore the space of Story Kits. We define a Story Kit as a themed set of characters, scenery, and animations designed to help inspire story ideas. Initially, we envisioned three ways in which a Story Kit could be presented to users: 1) in a folder in the Alice gallery of characters and objects (Users of Storytelling Alice construct the cast and setting for their stories by selecting objects from a gallery of pre-made 3D objects).; 2) as a Storytelling Alice program that contains a scene, characters, and an initial animation; or 3) as a combination of a Storytelling Alice program and a supplementary folder of content that users can add through the Storytelling Alice gallery. We chose to focus on Story Kits as a starting point for two reasons: 1. Girls in our user tests were often attracted to coherent sets of objects within the gallery. For example, they frequently selected the models and characters from Egypt, Japan, and the Amusement Park, the only coherent sets in the original Alice gallery. 2. Story Kits provide a low-cost way to experiment with different ideas. Making rapid, large-scale changes to the full Storytelling Alice gallery which contains more than 350 Inspiring Stories through Content 7 models was not feasible. Story Kits provided the opportunity to identify promising approaches by quickly developing and testing smaller sets of models and animations. Thirteen undergraduates who had prior experience with Alice participated in the Story Kits seminar. The undergraduate students worked in teams of three to four students to create and test a series of Story Kits. Over the course of the semester, we created and tested a total of 16 Story Kits in four rounds with each round taking two to three weeks. Each team functioned as a small research group; they were required to propose a mechanism for scaffolding story generation and describe how that mechanism would be realized in their Story Kit. Then, each team created the 3D geometry and textures for all characters and scenery elements in their Story Kit and animated their 3D models in Storytelling Alice. At the beginning of the next round, Story Kit creation teams were shuffled so that the undergraduate students were working with a different team on each Story Kit that they built. This approach of two-week long projects and shuffling teams for each project was inspired by the Building Virtual Worlds course (BVW, 2008). By completing four full rounds of development and testing, we hoped to facilitate the exploration of a wide variety of potential strategies for scaffolding story ideas. The motivation behind shuffling the groups for each round was to enable the Story Kit teams to more fully incorporate lessons learned from the previous rounds. By creating new groups we helped to ensure that each of the Story Kit team members had a unique perspective on the problem of inspiring story ideas, based on his or her experience with past kits. To enable evaluation of a new pair of Story Kits each week, we staggered the team deadlines. Participants Throughout the semester, a group of 10 local children came to Carnegie Mellon to participate as informants (Druin, 2002) in the design process. The children ranged in age from 10 Inspiring Stories through Content 8 to 15, 7 were female, and 6 were African-American. 4 attended public or private school and 6 were home-schooled. The weekly sessions with the participants were 1.5 hours long. During the first session, we explained that we were trying to understand how to build Story Kits that could help middle school students to find ideas for stories. We also introduced the participants to the basics of Alice, concentrating on the features that we felt would be useful in creating stories. During the subsequent Friday sessions, we asked the children to work in pairs to create a story using one of our Story Kits. Occasionally, because of absences or disagreements, children created stories individually. To focus our attention on each of the Story Kits, we instructed our participants to use only the content in the Story Kit to which they were assigned each week. To supplement the capabilities of the Story Kits, participants could use any of the standard Alice 2 animations. These include the ability to move and rotate characters or their body parts in space as well as say, think, and play sound animations that can be used to communicate the action of a story. While worked with a Story Kit, undergraduate Story Kit designers observed the process and recorded notes about what the participants did and said while creating their stories. To get the most objective information about the success of each Story Kit, we ensured that observers were not involved in the design or development of the Story Kit being tested. Observers recorded the process that participants went through in creating their stories, paying particular attention to suggestions of plot lines or character actions and the context in which those suggestions occurred. Further, the observers studied how engaged participants were in the process of creating a story through looking at cues such as the amount of off-topic conversation, verbal comments like “cool,” frustrated sighs, etc. At the end of each session, we asked the participants to offer suggestions on how to improve the Story Kits. The observations, user’s comments and Inspiring Stories through Content 9 suggestions, and the programs that the participants produced were used to provide insight into what attributes of a particular Story Kit contributed to its success of failure. By mixing the Story Kits teams for the next round, we helped to ensure that each of the teams had a more complete view of the strategies that had or had not been successful in the previous rounds.