Measuring Young Learners' Open-Ended Agent-Based Programming Practices With Learning Analytics

• In Event: Robotics and Social Agents

Abstract

Agent-based modeling (ABM) has been recognized as an important component of computational thinking and literacy. Agent-based programming (ABP) is the computational foundation of ABM. In this study, we propose a novel method to programmatically measure young learners' ABP practices in open-ended programming contexts with a 4-stage model. We present results from the online community of Turtle Universe, a new version of NetLogo designed for mobile platforms and younger learners in informal contexts. We draw on a dataset of 2,300 block-based and text-based projects shared by out-of-school, unsupervised learners. Our approach successfully measures learners’ open-ended ABP practices. Differences were found among projects of different stages and modalities. We discuss the implications of our approach and design of block-based ABP environments.

Authors

• John Chen, Northwestern University
  Presenting Author

• Uri J. Wilensky, Northwestern University
  Non-Presenting Author