Learning by self-explaining (e.g., justifying a solution step in a worked solution) is a powerful method when implemented appropriately. In this chapter, we first describe the historical roots of the self-explanation concept and we explain why we focus on principle-based self-explanations (i.e., interrelating domain principles such as mathematical theorems with problem cases). If students engage in this type of self-explanations they are likely to gain an understanding of domain principles (e.g., mathematical theorems) and their application. Such an understanding is a highly beneficial precondition for solving transfer problems in which the principles apply. However, it is necessary to consider several factors when implementing learning by self-explaining; otherwise this learning method’s potential is not fully exploited. Five such factors are discussed: self-explanation prompts that provide varying degrees of structure, self-explanation training interventions, working memory load induced by the basic learning arrangement, self-explaining correct as well as incorrect solutions, and the phase of skill acquisition. We also outline promising lines of further research. Finally, we summarize what practice recommendations can be provided according to the latest state-of-the-art.