The author of this paper reports on her doctoral work in community interpreting. She takes an interdisciplinary approach, as her particular viewpoint is sociological and discourse-analytical. After a brief presentation of discourse analysis, she focuses on methodological issues in a somewhat more practical vein. She underscores the fact that her approach is naturalistic, in that it involves observing people interact in the field rather than experimenting with them under controlled conditions, as is more common in cognitive studies. This requires gaining access to observable situations involving interpreters, as well as obtaining permission to record the relevant events. The author observes these issues, as well as the principles underlying the use of transcripts for discourse-analytical work, corpus size in such observational studies, how to record the interaction, and provides references for further reading.