Survey Design
Overview
Survey quality is mostly design quality. Poor questions produce confident, misleading numbers. Good design maximizes truth per minute of respondent time.
When to Use
- Customer, employee, or market questionnaires
- Tracking studies and pulse surveys
- Concept and pricing tests via survey
- Academic or program evaluation instruments
Core Practices
- Start from decisions and constructs, not a wishlist of questions
- Use simple, specific language; avoid double-barreled items
- Choose scales intentionally (and keep them consistent where comparing)
- Randomize or rotate sensitive order effects when needed
- Pilot for comprehension and timing
- Plan analysis before fielding
Principles
- Every question costs attention — earn it
- Leading and loaded wording poison results
- Mandatory long surveys increase junk data
- Incentives and channel choice affect who responds
Verification
- Each question maps to an analysis need
- Pilot feedback was incorporated
- Sampling and nonresponse risks are acknowledged