A measure of similarity of educational items has many applications in adaptive learning systems and can be useful also for teachers and content creators. We provide a thorough overview of approaches for measuring item similarity. We document the computation pipeline, explicitly highlighting many choices that have to be made in order to quantify item similarity. We also discuss methods for the analysis of similarity measures. For illustration and evaluation, we consider items from diverse domains, e.g., mathematics, programming, and language learning. Although there is no ultimate, universal approach to measuring item similarity, our overview leads to guidelines that facilitate computation of item similarity in practical applications.