Our first two experiments on adapting a high-structure course model to an essentially open-enrollment university produced negative or null results. Our third experiment, however, proved more successful: performance improved for all students, and a large achievement gap that impacted underrepresented minority students under traditional lecturing closed. Although the successful design included preclass preparation videos, intensive active learning in class, and weekly practice exams, student self-report data indicated that total study time decreased. Faculty who have the grit to experiment and persevere in making evidence-driven changes to their teaching can reduce the inequalities induced by economic and educational disadvantage.