
Computer programming is regarded as a difficult subject at both school and university. There have been a vast amount of studies with a focus on identifying students' difficulties, common errors and misconceptions in programming, and on the development and design of instructional techniques that could potentially help students overcome these difficulties. Nevertheless, there are few studies that explore students' performance in programming under the prism of self-regulation theory. To this end, the current study considers girls' and boys' calibration and how it is related with their performance in programming, self-evaluation, and self-efficacy in computer science. Calibration is a measure of the accuracy with which people assess their confidence in their own performance. The results of our study suggest that boys feel significantly more efficacious in computer science than girls, as well as make significantly more accurate predictions (better calibrated) of their programming performance than girls. The implications of these findings for the current education practices are outlined and discussed.