Does the modality of instructional messages affect learning? How does it affect different media? In this paper, I offer an answer to these questions by first proposing a theoretical framework from which effective instructional methods can be derived. Then, I report a set of studies where one method, the modality principle, was tested across different media and found to produce significant learning benefits. Based on the findings and those of a recent meta-analysis of modality effects, I suggest future research directions to examine method and media contributions to learning.