A study of doctor-patient communication offers convincing arguments for using multiple methods of data collection incorporating both meanings and practices. Multiple realities emerged clearly that would have remained invisible had only one data source been used. Two case studies illustrate how four factors interact to produce different versions of reality: setting, participants, time, and forms of data recording. The author discusses the apparent markedly different realities of doctors and patients, and the researcher's role in synthesizing these multiple accounts. Only by using multiple methods can attention be paid to the central tensions, the gaps and white spaces, and the discrepancies and misunderstandings that are so important in understanding human interaction. However, this approach is labor- and time-intensive, and requires skilled, experienced researchers.