CRII: CHS: Automatically Praising Learning Process to Promote the Growth Mindset in Computer Science

Northwestern University, Evanston IL

Investigators

Abstract

There is a pressing need to train large numbers of computer scientists to meet the demands of our nation's economy, but many students struggle in introductory programming courses. Recent studies show that these courses often promote the fixed mindset, or the belief that programming aptitude is an inborn trait. Psychology research shows that students with the fixed mindset view mistakes as indications of low ability and perform poorly in the face of challenge. In contrast, students with the growth mindset believe that programming aptitude is malleable and excel when challenged. This project aims to design, build, and evaluate programming tools that help students develop the growth mindset by automatically detecting and praising good learning behaviors as students write code. This research will contribute scientific knowledge about the growth mindset in the domain of computer science and provide insights about the process of learning to program. The project team will deploy the tools to hundreds of students at their own university and release them for free online for any student or teacher to use. If successful, this intervention has the potential to improve the experiences, skills, and diversity of students who successfully complete programming courses and go on to participate in employment and research in STEM fields. This project aims to develop a new growth mindset intervention that leverages the programming environment by using artificial intelligence techniques to automatically detect and praise good learning processes in real time. Programming environments provide a unique opportunity to track and understand student learning behaviors, and offer a scalable environment for praising good practices automatically. By exposing and praising the learning process, this intervention will teach students to attribute their successes and failures to malleable learning processes, rather than an innate aptitude for computer science. This research will be conducted in two phases. First, the project team will develop heuristics that detect good learning processes using behavioral log data, leveraging the computer science education literature and studying the behavior of fixed and growth mindset students to identify good processes. Second, the team will iteratively design and build a programming environment extension that uses the validated heuristics to automatically detect and praise good learning process, and evaluate this intervention through a controlled ten-week study with university students. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.