3.3.2, 3.3.4.2 - Collecting relevant, good-quality field evidence

Observation and usable raw records

Raw records preserve what was observed and how it was obtained, before an interpretation is imposed.

An observation needs enough information to be usable later: site or participant code, variable, value or response, units or category definitions, date/time, relevant conditions and collection method. Record readings at the time rather than reconstructing them from memory. Keep observation separate from inference: “three boarded entrances visible” is an observation; “the area is declining” is an interpretation that needs more evidence.

Consider an original fictional collection log:

P2 is not a zero: the absence of a completed count is different from observing nobody. P3 has a later window and changed weather, so direct comparison needs that qualification. The log makes those limitations visible. Inventing a count for P2 would create false evidence, not repair the missing observation.

Consistency supports comparison. Agree what counts as a pedestrian crossing, how to handle someone turning back and where counting begins. For physical measurements, use a consistent procedure, units and recording precision. Repeat measurements when appropriate, and retain unexpected readings while investigating possible errors; do not delete them only because they disagree with a prediction.

Primary and secondary evidence provenance

The origin of evidence determines how it can be used and what limitations need examination.

Primary evidence is collected by the candidate, individually or as part of a permitted group field investigation. Secondary evidence comes from another source, including raw measurements made by another cohort or data collected previously. Independently select and source relevant secondary material; retain the author/organisation, location, period, variables, units and method information available.

Complete and obtain approval of the CRF proposal before collecting NEA primary data. AQA's supplementary guide distinguishes earlier practice data from that primary collection. Independently deciding to use an earlier dataset can make it a secondary source; its origin should remain clear. An external provider's dataset is not automatically the candidate's own primary evidence because it concerns the same place.

For permitted joint collection, record who helped and how the procedure was performed. Each candidate must independently choose how the evidence is used and write their own account. A shared raw table is compatible with independent analysis; copied reasoning is not. Secondary sourcing does not become a group task just because primary measurements were shared.