The main goals of this study were to test whether multiple representations, such as diagrams and equations, per se help to acquire conceptual understanding in probability, and to investigate whether learners need instructional support to utilize the potentials of multiple representations. The authors conducted an experimental study with 8 conditions in which high school students (N = 170) studied worked examples from probability. The authors varied the type and number of representations and the availability of 2 support procedures: (a) a relating aid that used color codes and flashing to help learners see which elements in different representations corresponded to each other on a surface level and (b) self-explanations prompts to ensure that learners integrate corresponding parts in different representations on a structural level. The authors found that multiple representations per se did not foster conceptual understanding, however, both support procedures enhanced it. Yet, the self-explanation prompts did not only foster conceptual understanding by eliciting elaborations directed to domains principles but also incorrect elaborations that hindered the acquisition of procedural knowledge. Hence, self-explanation prompts are an instructional support procedure that can have conflicting effects on learning outcomes. (Contains 3 figures and 7 tables.)