Survey bias
Build a sample and questionnaire that answer the decision at hand.
Quick learn
A fictional grocery chain emails its loyalty list to estimate demand for an evening pickup service. The survey asks whether people want a “convenient, time-saving service”. Eighty percent say yes.
Deep dive
Coverage bias appears when some of the target population cannot enter the sample. Nonresponse bias depends on who responds. Leading wording can change answers even in a well-drawn sample. Report the recruitment and full question wording.
Worked decision
Recruit across stores and time blocks, include non-members, and ask how often respondents would use pickup at a specified window and fee. Pilot the question in the languages shoppers use.
When the decision changes
Collect observed pickup trials alongside stated intent. If a small intercept sample undersamples late shoppers, use planned recruitment blocks and compare results by block before pooling.