How do you design a good survey that gives reliable data?

Surveys look easy and are easy to get wrong. Let's collect best practices.

Key principles:

• Ask one thing per question — avoid "double-barrelled" questions.
• Use neutral wording; don't lead people to an answer.
• Pilot it with a few people first to catch confusing questions.
• Keep it as short as honestly possible — long surveys lose respondents.
• Consider language: offering it in Somali and English may improve response quality.

What survey mistakes have you learned to avoid? Share so others don't repeat them.