Learning programming, syntax errors and institution-specific factors

Abstract

Learning programming is a road that is paved with mistakes. Initially, novices are bound to write code with syntactic mistakes, but after a while semantic mistakes take a larger role in the novice programmers' lives.Researchers who wish to understand that road are increasingly using data recorded from students' programming processes. Such data can be used to draw inferences on the typical errors, and on how students approach fixing them. At the same time, if the lens that is used to analyze such data is used only from one angle, the view is likely to be narrow.In this work, we replicate a previous multi-institutional study by Brown et al. [5]. That study used a large scale programming process data repository to analyze mistakes that novices make while learning programming. In our single institution replication of that study, we use data collected from approximately 800 students. We investigate the frequency, time required to fix, and the development of mistakes through the semester. We contrast our findings from our single institution with the multi-institutional study, and show that whilst the data collection tools and the research methodology are the same, the results can differ solely due to how the course is conducted.