Supporting the contention that using experimental research designs will facilitate sound decisions about how to teach the composing process, this paper first describes briefly the range of research methods available for investigating the writing process. Next the paper discusses the advantages of conducting experimental research, some of which are that it (1) encourages careful planning, close attention to classroom procedures, and systematic evaluation; (2) is a better test of the effectiveness of teaching methods; and (3) promotes a clearer understanding of research methodology. Finally, the paper illustrates these advantages by summarizing a "true" experiment (one whose components are an hypothesis, assumptions, a research design, randomized subjects, a treatment, and a measurement procedure) to learn whether dictation/word processing systems required a composing process somewhat different from that of writing. (RL)