Abstract:

Invention activities are carefully designed problem-solving tasks in which learners are asked to invent solutions to unfamiliar problems prior to being taught the canonical solutions. Invention activities are typically used in the classroom setting. As online education becomes increasingly common, there is a need to adapt Invention activities to the asynchronous nature and facilitate their delivery and analysis in a larger scale. We start by focusing on the invention process itself and its outcomes based on a case study in which we analyze video recordings we collected of several students who worked on these activities in pairs as part of an introductory undergraduate data science course. We discuss lessons learned and implications for the design of asynchronous Data science Invention activities. Then we focus on facilitating these activities at scale based on a second case study we ran in the following year in which we test the delivery and submission of the activities, and present analysis of the activities using a dedicated framework. The framework serves as an intelligent submission system to support scalability while also providing instant personalized feedback to the students to address challenges raised from the asynchronous nature of the activities. We use the framework to analyze the solutions and validate the efficacy of the activities at scale.