Therefore, estimates from the survey might be sensitive to systematic errors because respondents might differ from non-respondents in significant ways.