Reading skills can predict the programming performance of novices: an eye-tracking study

Abstract

Due to the character of programming languages, reading ability may have more impact on learning to program than on learning in other subjects. This paper describes an exploratory study of the relationship between reading skills, as perceived through eye tracking, and the ability to program. An empirical investigation into this relationship determined that students with inadequate reading skills are at risk of failing at introductory programming. As an explanation for the effect of reading ability on learning to program, we argue that a programming language is a special high-level written language and that using it requires high levels of comprehension, inferencing, selective attention, organising and reflecting. As a result, a student’s reading ability will have a considerable effect on learning to program. Lack of reading skills may therefore be a factor that affect students’ ability to learn to program. Eye tracking can expose reading skills and, therefore, be used to identify at-risk introductory programming students. The practical contribution of this research is the demonstration of how eye tracking can reveal reading problems among programming students. We relate these reading problems to their programming performance, providing a theoretical account of the connection. The results suggest that efforts to improve reading skills could have a positive impact on learning to program.