This article investigates why students reported liking a student-driven learning design better than a highly guided design despite equivalent gains in knowledge assessments in both conditions. We created two learning designs based on the distinction in the literature between student-driven and teacher-led approaches. One teacher assigned each of her two 5th-grade classes to one design (n = 52); both designs addressed the mathematical concepts for the same amount of time. Data were collected using written assessments, surveys, and video. Students in both classes improved equivalently on assessments. On surveys, students in the student-driven condition were significantly more positive about learning. Video was coded to examine why students were more positive in the student-driven design. This analysis showed students engaging more frequently with data in discussing strategies, questioning peers, and aligning outcomes with prior experiences. We trace the association between students' positive responses to learning and richer engagement in mathematical practices to specific features of the student-driven design. Furthermore, we reject competing explanations (e.g., amount of off-task behavior or adult intervention). We conclude that designing learning opportunities that promote mathematical practices affords opportunities to cultivate students' disciplinary interest. We discuss implications for teachers and curriculum designers who are responding to new mathematics standards.