Noting some of the major problems with current design experiments in education, a project has been undertaken with the long-term goal of constructing a systematic science of how to design educational environments so that new technologies can be introduced successfully. This paper outlines several factors considered to be critical in developing a methodology for conducting design experiments and provides an example of a proposed application of the methodology to the development of a multimedia teaching unit about the seasons which would incorporate the film "The Voyage of the Mimi 2," associated computer programs, a program for teaching students how to construct tables and graph data, a computer network, and application programs such as word processors and drawing programs. It is noted that the evaluation of the unit, through its reliance on multiple data collection methods--including pre- and post-tests, structured interviews, classroom observations, teachers' comments, and follow-up studies--would avoid the shortcomings of current design experiments and would help determine the form that a design theory should take. Finally, the initial phases in constructing such a theory--i.e., identifying all relevant independent and dependent variables by which the success or failure of an innovation can be measured and specifying how these variables interact--are discussed. Factors affecting the success of technology in education are presented in two tables. (GL)