Capturing Kindergarteners' Computational Thinking Through Commercial Toy-Centered Task and Assessment Development

• In Event: Fostering Young Children's Computational Thinking

Abstract

A number of commercial toys have appeared that espouse their ability to support computational thinking (CT) for young children. Whether and how these do so when children actually play with these toys remain to be determined. The purpose of this presentation is to report on our early efforts to elicit and document CT and associated behaviors in kindergarten-aged children based on play with these toys.

To do this, we are engaging in cycles of design-based research (Cobb et al, 2003) to iteratively develop toy-centered tasks for kindergarteners based on CS standards (e.g., Computer Science Teachers Association, 2017). In parallel, we are developing assessment tasks using evidence-centered design (Mislevy & Haertel, 2006) that can elicit children’s knowledge, skills, and abilities with respect to different aspects of CT, such as algorithmic thinking or debugging.

We have obtained video data to inform iterations of both task and assessment activities (Derry et al, 2010). This includes children working on the tasks in groups at two public (n = 39) and pairs in two private (n = 23) kindergarten classrooms. Sixty-one children across all settings participated in the pilot of assessment items through videotaped interviews.

One common feature in the design of toys, evidenced by implicit use of spatial grid systems, was a latent assumption of body syntonicity (Papert, 1980) – an ease with which mappings are made from one’s own body and its movement in space to the moving agent as an object to think with. However, we are finding that body syntonicity was not always intuitive for children. Specifically, programming toys to move laterally (i.e., translation) was often conflated with rotation and turning commands. Children gave instructions to move a toy to an adjacent square and found that their commands yielded only a change in orientation and not a change in position. Moreover, in an assessment task to examine how initial toy orientation affected mastery of movement commands (forward, back, turn right, turn left), right and left were often used incorrectly when the relative perspective of the toy was opposite of the child. This finding problematizes the intuition that body syntonicity is by default natural for young children.

Another behavior we observed involved children modifying incorrect movement codes by adding to the end of the existing code even if a more efficient solution was possible by changing codes that were already given. We refer to this tendency to add to the end of existing code as the “appendment” strategy of code modification as opposed to “amendment” where some inner portion of code was edited. Still, using this appendment strategy, children were able to debug the code and make solutions that worked. In the presentation, we discuss this and other debugging strategies exhibited by kindergarteners.

Authors

• Victor R. Lee, Stanford University

• Jody E. Clarke-Midura, Utah State University

• Jessica F. Shumway, Utah State University

• Joseph Kozlowski, Utah State University

• Lise E. Welch, Utah State University

• Hannah Evans, Utah State University