Interaction-free measurement is introduced through a set of problems suitable for undergraduates. Both the original scheme suggested by Elitzur and Vaidman and an improved one proposed by Kwiat et al. are considered. Theoretical predictions are compared to experimental data.