Automatic Formative Assessment in Computer Science: Guidance to Model-Driven Design

Abstract

Adaptive online learning can facilitate students' support by responding immediately to the user's interactions. Good feedback to students helps closing the gap between actual and desired performance. In this paper we analyze how to introduce online adaptive formative learning in Computer Science, a discipline with well documented challenges that are hard to tackle with traditional classroom methods. Specifically, we developed illustrative learning items teaching Model-Driven Design and implemented them in an online system that implements a model for automatic formative assessment developed by University of Torino. The model takes advantage of an automatic assessment system initially designed for STEM disciplines, then adopted for teaching languages and other disciplines too. The key features of the adaptive model supported by the online system are algorithmic questions, availability, contextualization, immediate feedback, interactive feedback, and open answers. These features are portable across subject domains, so the system can be adapted to include new subjects. We chose MDD because it is a topic of Computer Science education connected with Computational Thinking, software design, and formal methods, which are three of the core areas in need of enhanced support.