There is one thing the Computer Science Education researchers seem to agree: programming is a difficult skill to learn. Educational technology can potentially solve a number of difficulties associated with programming and computer science education by automating assessment, providing immediate feedback and by gamifying the learning process. Still, there are two very important issues to solve regarding the use of technology: what tools to use, and how to apply them?

In this thesis, I present a model for successfully adapting educational technology to computer science and programming courses. The model is based on several years of studies conducted while developing and utilizing an exercise-based educational tool in various courses. The focus of the model is in improving student performance, measured by two easily quantifiable factors: the pass rate of the course and the average grade obtained from the course.

The final model consists of five features that need to be considered in order to adapt technology effectively into a computer science course: active learning and continuous assessment, heterogeneous exercise types, electronic examination, tutorial-based learning, and continuous feedback cycle. Additionally, I recommend that student mentoring is provided and cognitive load of adapting the tools considered when applying the model. The features are classified as core components, supportive components or evaluation components based on their role in the complete model.

Based on the results, it seems that adapting the complete model can increase the pass rate statistically significantly and provide higher grades when compared with a “traditional” programming course. The results also indicate that although adapting the model partially can create some improvements to the performance, all features are required for the full effect to take place.

Naturally, there are some limits in the model. First, I do not consider it as the only possible model for adapting educational technology into programming or computer science courses. Second, there are various other factors in addition to students’ performance for creating a satisfying learning experience that need to be considered when refactoring courses. Still, the model presented can provide significantly better results, and as such, it works as a base for future improvements in computer science education.