The purpose of this research was to examine the effects of instruction on students' generation of questions when reading multiple documents. Participants were 151 undergraduate students who were randomly assigned to one of two conditions: a treatment condition that received question generation instruction and a control condition that did not receive question generation instruction. All participants read four primary-source documents and were asked to generate up to 10 questions based on those documents. Results indicate that students generated significantly fewer higher-order integration questions without instructional support. These results are replicated with an alternative document set to demonstrate that the effects found were not due to the specific historical topic.