At the heart of contemporary classroom research is the question of whose accounts of classroom activity are priveleged for the purpose of understanding learning processes in such settings. Classroom reserch has been greatly assisted by the emergence of a varieyt of technologies to support both data collection and analysis. This paper contrasts the techniques and the assumptions by which interpretations and meanings are constructed by contemporary classroom researchers. Specifically, comparison is made between two approaches to classroom research: "Validity by consensus" in which transcripts of classroom dialogue and activity are analysed by several researchers and consensus as to the "meaning" of an episode is sought through the synthesis of the many interpretations; and, "Complementary accounts" in which researchers interpret the documented interaction from several distinct theoretical perspectives, and the goal is diversity rather than consensus, where each interpretation is accorded parity of status, together with the student's reconstruction of the episode, prompted by the video record of classroom events. In the first approach, the matter to be interpreted is the text of transcribed classroom dialogue and activity, while in the second, analysis is undertaken of integrated data sets combining videotape, dialogue transcripts, and video-stimulated reconstructive interviews.