Teaching Computational Thinking through Multi-Agent Simulation: Increasing Recruitment, Retention, and Relevance of Undergraduate Computer Science

Middlebury College, Middlebury VT

Investigators

Abstract

The number of students studying computer science and computational thinking has been declining at an alarming rate. New approaches to attract students are needed. This project is developing an introductory course that presents computational thinking in a broad context. Realistic problems from disciplines such as biology and economics are used to show relevance. The approach emphasizes multi-agent simulation using the NetLogo programming language. Multi-agent computer simulation is based on the interactions of numerous intelligent agents with each other and with an environment. For example, individual consumers are represented as interacting agents to model a complex economic system. Although the course is designed as an introduction to computational thinking for science and economics majors, the project is investigating how the course can be used as an alternative CS1 course for computer science majors. Assessment plans include measuring student learning as well as monitoring recruitment of new students to computing. Expected outcomes include course materials and faculty training materials. Results will be disseminated at regional and national computer science education conferences.