Investigating the Utility of Notional Machine Instruction in an Introductory Programming Lesson

Learning to program involves understanding how the computer executes programs. Code tracing is a simulation of the steps the computer takes to execute a program. The notional machine is an abstract representation of this process. Students form mental models of the notional machine when learning to code trace, but these mental models can be inaccurate or contain misconceptions. We experimentally investigated the effect on learning of using an explicit notional machine or not in a code-tracing lesson for novice programmers (N = 48). We created two versions of a tutoring system, one with a notional machine and one without, using the Cognitive Tutor Authoring Tools framework. The tutors included video lessons and self-explanation prompts to encourage participant engagement. We measured learning as the difference in scores from pretest to posttest, adjusted by prior knowledge. Learning increased overall, but there was no significant difference in learning between groups.