If research into writing, particularly on computers as writing tools, is to assist teachers to improve writing pedagogy, the emphasis should be toward conducting naturalistic and context-sensitive studies. Early studies involved traditional experimental research instruments and looked not at the learning process but at the written product, measuring number of words, errors, number of revisions and quality of product rather than evaluating focus, organization and the ability to use arguments, details and examples effectively. When asking if student writing is instantly transformed via the computer's magic wand the more relevant question is not whether the writing is good or bad, but rather whether students have become more engaged by the writing process and acquired more sophisticated techniques for composing, revising and editing. Better instruments for evaluating and comparing texts must be devised if research results are to be significant. Although the important relationships between context and writing have gone unexplored until fairly recently, they are increasingly recognized as an important element; context-sensitive classroom-based research can provide important information for developing teaching strategies and hopefully shed light on the relationship to the students' progress as learners. (NH)