Two widely disseminated approaches impose reductive boundaries on ethnographic research by privileging one context of meaning over other essential contexts. The first, emphasizing statistical validity, privileges the research community by recommending that the ethnographer's data analysis via coding agree with that of other raters from the research community. The second asserts that the ethnographer who comes closest to validity comes closest to presenting only the subject's point of view. Ethnography, however, comprises four essential, overlapping contexts: the phenomenal context (that which is observed/recorded), the site's cultural context (the subjects' outlook), the research community context, and the researcher's interior context, shaped by experience and education. Each of the four vantages has dominating tendencies, but if one does dominate to the exclusion of others, the reductive result is data-centered, thin description; subjects-centered groupthink; research community-centered groupthink; or researchercentered solipsism. Although all contexts of meaning are important, none should fully eclipse the others.

Previous research in computer-mediated communication in both the classroom and the workplace has found that patterns of interaction may differ between individuals communicating face-to-face versus communicating via a computer network. This present study, using a case study methodology, sought to analyze and compare the language of groups of business writing students as they communicated both face-to-face and on a real-time computer network. The study found that during network meetings, participation was more equal, responses tended to be more substantive and text specific, and students were more willing to offer direction than during face-to-face meetings. In addition, students reported a more positive evaluation of their network sessions.

What is the role of interaction, or, more generally, orality, in multiple-audience analysis and adaptation? How does orality relate to literacy in the evolution of corporate documents? A qualitative study of how seven engineers in two divi sions of a large corporation wrote for multiple audiences revealed that, in the more rhetorically successful cases observed, interaction was the central means of analyzing and adapting discourse to multiple audiences, fulfilling rhetorical and social goals, and building and sustaining a corporate culture; and orality was more potent than literacy in the engineers'composing behavior and the au diences' acceptance of the engineers' ideas and documents.