The efficacy of courses that mix face-to-face and online instruction, such as blended, hybrid, flipped, and inverted courses, is contested in the literature. Some studies find that they improved learning outcomes and some do not. We argue that these unreliable results are due to inconsistent definitions of these courses. To address this problem, we propose the Mixed Instructional eXperience (MIX) taxonomy to define hybrid, blended, flipped, and inverted based on two dimensions. To test the usefulness of the taxonomy to organize the literature, we reclassified research using the taxonomy. The analysis of the literature after reclassification revealed themes that illuminate how mixing face-to-face and online instruction affects learning. These findings validate the taxonomy as a useful tool for classifying literature and further knowledge in this field.