The aim of the paper is to provide guidance on conducting participant and non-participant observation studies of information behaviour. Examines lessons learned during non-participant observation of hospital pharmacists, and participant observation with dependent older people living in their own homes. Describes the methods used in both studies, and discusses the ethical issues involved in gaining access to the subjects. In the hospital setting, professional affiliation between the researcher and the subjects (6 pharmacists) made access easier to obtain. In the home care setting, access to subjects (7 clients) for participant observation (as a care worker) was more difficult, as was withdrawal from the field study. In both studies, the observation element was triangulated with survey data. Both studies indicated the fundamental need for trust between the observer and the research subjects. In some situations, professional relations offer instant access and trust, whereas in closed and sensitive situations such as social care, time is required to build up trust. With participant observation, that trust should not be damaged by withdrawal of the researcher from the research setting.