The founders of the field did not do fieldwork. Nineteenth-century writers on religion built their comparative theories from reports sent home by missionaries, administrators and travellers, and the material arrived already shaped by the interests of the people who collected it. The results were confident, systematic and frequently wrong.
The methodological revolution was participant observation: living in a community for an extended period, learning the language, and describing practice as it is actually conducted rather than as a questionnaire records it. It transformed the evidence base and it remains the discipline's distinctive instrument.
It also brought its own difficulties, which the field turned on itself in the 1980s. A critique gathered around the writing of ethnography argued that a monograph is a composed text with a narrator, that the authority of the omniscient observer conceals the negotiation and misunderstanding through which the material was actually obtained, and that the people described had no voice in how they appeared.
The practical consequence has been procedural. Consent, community review of what is published, restrictions on restricted knowledge, and co-authorship with the people studied are now normal expectations rather than courtesies, and the older monographs are read with their conditions of production in view.
For the study of religion specifically that matters most where knowledge is deliberately withheld within a tradition, since publishing it is not neutral even when it is accurate.