Arrange the stages of conducting a survey for research. A. Pilot testing of…

Concept

A survey collects data through a standardised instrument, so its stages are fixed by a single dependency rule: every stage produces the input that the next stage consumes. An instrument must exist before it can be trialled, it must be trialled and corrected before it goes to the full sample, and responses must be gathered before they can be interpreted.

Applying the rule to these four stages

1. Designing the questionnaire (B) — the items, wording, response scales and skip logic are written. This creates the instrument, so no other stage can precede it.

2. Pilot testing of the survey (A) — the draft instrument is administered to a small sub-sample to expose ambiguous items, faulty routing and length problems while revision is still cheap.

3. Distributing the questionnaire (C) — the corrected instrument is administered to the full target sample and the responses are gathered.

4. Analyzing the responses (D) — the collected data are coded, tabulated and interpreted to answer the research question.

The stage order is therefore B, A, C and D.

Cross-check

• Pilot testing needs a drafted questionnaire, so designing must come before pilot testing.

• Distribution needs an instrument already corrected by the pilot, so pilot testing must come before distribution.

• Analysis needs responses that have actually been returned, so distribution must come before analysis.

• An arrangement that breaks any one of these dependencies describes an impossible workflow — trialling an instrument that has not been written, or interpreting responses that have not yet been collected.