MKT.12:4.6 - Obtain and interpret enough evidence

Choose cases that can distinguish the live explanations, including access failures or non-users when their absence would bias the answer. An initial purposive set can discover differences. A population estimate needs a defensible population and sampling design; a convenient set of willing customers does not acquire representativeness by growing larger.

Arrange permitted access, understandable participation and competent inquiry. Pilot consequential questions on a small attainable case to find ambiguity before wider collection. Keep the question wording, relevant context and source of observations recoverable in the ordinary working material. Retain only information needed for the stated use under the applicable conditions. Obtain specialist advice where access, rights or a statistical claim depends on it.

During analysis, separate observed events, participant reports and your inferences. Compare cases and examine contrary observations. Ask whether a difference could come from the way information was obtained: selective response, changed wording, interviewer prompting, a different period or an incentive to give a favourable answer. A plausible explanation is useful as a hypothesis when it guides the next discriminating move; it need not be inflated into a general law.