A comparison of different types of learning activities in a mobile Python tutor

Programming (i.e. coding) is becoming one of the skills expected for successful careers in the knowledge economy1, and is being taught at all levels, including primary and secondary schools. Programming skills are difficult to acquire, as the student needs to learn the specific programming language and many related concepts to write good programs. We present PyKinetic, a mobile tutor for Python that serves as a complement to traditional courses. The overall goal of our project is to design learning activities that maximize learning. In this paper, we present several types of learning activities designed for PyKinetic. The first version of the tutor implemented Parsons problems with incomplete lines, which support code-understanding and code-writing skills. The second version of PyKinetic included various types of activities aimed at code-tracing and code-writing skills. The results of two studies we conducted show that Parsons problems are beneficial for novices, while advanced students benefitted more from learning activities which required them to identify and fix incorrect lines of code