Students in two iterations of a statistics course were required to develop work-related research questions that could be analyzed by the statistical techniques discussed in class. They were also expected to present the question and other research design components in a manner that could be comprehended by statistical novices. To provide sufficient class time for the groups to begin work on their assignments, the course's second iteration substituted Kahn Academy videos for a portion of the lecture material. However, observations of the students' behaviors and qualitative analysis of the responses to open-ended course evaluation questions indicate that efforts to encourage self-directed learning were undermined by an allegiance to stimulus-response learning, i.e. a preference for courses in which students are told what they need to know. Since the preference is based on a desire to maximize the grades generated by a given investment of time, the impending iteration of the course incentivizes self-directed learning by incorporating into the course grade calculation the students' contributions to their group's efforts and participation in class discussions. The revised grading structure is intended to encourage self-directed learning in which students teach and learn from one another and thereby engage in deep learning.