Providing test questions after an initial study phase is a common instructional technique. In theory, questions that require higher-level (deep) processing should be more beneficial than those that require lower-level (shallow) processing. However, empirical evidence on the matter is inconsistent. To shed light on two potential reasons for these inconsistencies, we conducted an experiment (N = 81) to analyze the role of (a) the share of information to which test questions direct learners and (b) the learning activities performed in the initial study phase with respect to the effects between higher- and lower-level test questions. We found that both the share of focused information and the initial learning activities moderated the effects between higher- and lower-level test questions.