A large sample is not representative when certain groups do not respond. A nonresponse plan establishes who is missing, how they will be approached and how uncertainty will be disclosed if the gap remains.
Define the reference population
Record who should have had a chance to participate and which sampling frame was used. An open link does not establish a known sampling base.
Compare respondents and nonrespondents
Use available, safe characteristics such as region, age band or user type. Identify where participation differs systematically.
Reduce friction
Make the survey short, mobile friendly, accessible and clear about the time required. Try different times, channels and reminders without pressuring people.
Target the gaps
Use additional outreach or a different collection method for underrepresented groups. Increasing responses from people already willing to take part does not correct bias.
Run a sensitivity analysis
Apply weights carefully and examine how much the conclusion changes under reasonable assumptions about people who did not respond. State the limitation clearly.
This framework is an original editorial methodology developed by DigitalNow.