Revisiting the Confidence Gap in University-Level Programming Courses

Abstract

<p>The confidence gap, that is, the difference in self-assessed confidence between genders regarding the ability to succeed in a given task, has been previously reported and studied in the context of computer science education. In this paper, we investigate the prevalence of the confidence gap during subsequent introductory and advanced Python programming courses - attended by a range of students from CS majors to non-university students taking the courses as free MOOCs - from a variety of perspectives, including student perceptions of what constitutes as 'good' or a 'bad' grade, what they predict as their own grade at the start of the course, and previous exposure to programming in various contexts. Our results provide both an updated snapshot into the evolving confidence gap, and additional data about students' beliefs on how well they are likely to perform in university-level programming courses targeted at first-year students.<br></p>