You need data on the financial behavior (banking, loans, etc.) of mid-Americans to solve a particular problem your firm is experiencing. You design a questionnaire that you pretest at a local bank officers' club meeting. You adjust the questionnaire and send it to a stratified sample of mid-Americans in a large Midwestern city. Your strata (see assume sub-samples) are defined geographically and consist of three areas. These are inner city, close outskirts, and suburbs.
What possible things could go wrong?